sneak/smallwebwaf
1
0
Fork 0
Code Issues 43 Pull Requests 2 Actions 44 Packages Projects Releases Wiki Activity

DNS blocklists asked in the background, verdicts kept (closes #104)
check / check (push) Waiting to run
Details

Browse Source 
Zones in SWWAF_DNSBL_ZONES are asked about each client in the background,
through SWWAF_DNSBL_RESOLVER or the host's resolver; no request waits.
Verdicts last SWWAF_REPUTATION_CACHE_TTL, kept in reputation.json.
SWWAF_REPUTATION_ACTION (limit:25) denies, limits or logs a listed
client; each zone listing it raises reputation_hit. A failed query gives
no verdict, raises source_failure, and pauses the zone a minute. A
zone's key, its first label under dq.spamhaus.net, is masked everywhere
but reputation.json. Zones compare without regard to case.

Judgement call: answers in 127.255.255.0/24 or outside 127.0.0.0/8 are failures.
Judgement call: the minute's pause after a failure; at most 1,000 queries at once.
Judgement call: one zone given with two keys stops the start as listed twice.
Rule suppressed: paralleltest on the DNSBL tests (Go's resolver shares state across synctest bubbles), funlen on the test of every logged setting.

Model: opus-5-5
This commit is contained in:
clawbot
2026-10-07 18:47:20 +00:00
parent 2b8c98ba1f
commit 15868fe491
18 changed files with 2330 additions and 236 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+255 -140
View File
@@ -26,33 +26,36 @@ Slack and ntfy, and remote log sending. So is the stage after that: the AS
number and country of every client, looked up through GeoJS or in the IPinfo
Lite database file, the byte limits, the biased thresholds, lower limits for the
AS numbers and countries you list, and the anomaly thresholds, alerts for
unusual traffic that refuse nothing. So is the first part of the stage after
that: the blocklists you name by URL, which it fetches and keeps, and a file of
AS numbers' percentages, fetched the same way. `smallwebwaf` passes each request
to the app and the app's answer back, unchanged, within its timeouts and size
limits, works out each client's address, looks up its AS number and country
unless you switch that off, bans a client that sends too many requests or too
many bytes, not counting those for the paths you choose, with lower limits for
the clients of the AS numbers and countries you list, refuses a client that
comes from a country you refuse or from a network you refuse, refuses, limits or
only notes a client a blocklist you name lists, lets the networks you choose
through, checks each request against the rule files and bans a client whose
request is a clear sign of attack, keeps its bans, each client's counters and
history, GeoJS's answers and the last good copy of each list it fetches in JSON
files across restarts, takes in your edits of those files, such as a ban you
make, keep or lift, and of the rule files while it runs, writes a JSON log line
for every request, sends its log lines to a syslog server too if you name one,
sends an alert to a webhook, to Slack and to ntfy, each if you name one, for
each ban it makes or makes permanent, for traffic over an anomaly threshold you
set, for a client a blocklist lists, for GeoJS failing or a list it cannot
fetch, for a rule file or state file with an error and for a replacement of the
lookup database it cannot read, serves Prometheus metrics to a scraper that
holds the metrics token, lets an admin who holds the admin token list, add and
lift bans and ask what it knows of a client, and in `observe` mode passes on the
requests it would refuse, logging what it would have done with them. It comes as
the image the app's own image is built on. The rest of the design comes after
that, in the order of the build order in [`SPEC.md`](SPEC.md). The survey of
existing tools that led to the design is in [`EVALUATION.md`](EVALUATION.md).
unusual traffic that refuse nothing. So are the first two parts of the stage
after that: the blocklists you name by URL, which it fetches and keeps, with a
file of AS numbers' percentages fetched the same way, and the DNS blocklists
(DNSBL zones), which it asks about each client in the background. `smallwebwaf`
passes each request to the app and the app's answer back, unchanged, within its
timeouts and size limits, works out each client's address, looks up its AS
number and country unless you switch that off, bans a client that sends too many
requests or too many bytes, not counting those for the paths you choose, with
lower limits for the clients of the AS numbers and countries you list, refuses a
client that comes from a country you refuse or from a network you refuse,
refuses, limits or only notes a client a blocklist or a DNSBL zone you name
lists, lets the networks you choose through, checks each request against the
rule files and bans a client whose request is a clear sign of attack, keeps its
bans, each client's counters and history, GeoJS's answers, the last good copy of
each list it fetches and the DNSBL zones' verdicts in JSON files across
restarts, takes in your edits of those files, such as a ban you make, keep or
lift, and of the rule files while it runs, writes a JSON log line for every
request, sends its log lines to a syslog server too if you name one, sends an
alert to a webhook, to Slack and to ntfy, each if you name one, for each ban it
makes or makes permanent, for traffic over an anomaly threshold you set, for a
client a blocklist or a DNSBL zone lists, for GeoJS failing, a list it cannot
fetch or a DNSBL zone that fails or refuses a query, for a rule file or state
file with an error and for a replacement of the lookup database it cannot read,
serves Prometheus metrics to a scraper that holds the metrics token, lets an
admin who holds the admin token list, add and lift bans and ask what it knows of
a client, and in `observe` mode passes on the requests it would refuse, logging
what it would have done with them. It comes as the image the app's own image is
built on. The rest of the design comes after that, in the order of the build
order in [`SPEC.md`](SPEC.md). The survey of existing tools that led to the
design is in [`EVALUATION.md`](EVALUATION.md).
## Getting started
@@ -206,35 +209,48 @@ in `bin/state` unless `SWWAF_STATE_DIR` is set, and the default rule file of
limit, the lowest of its percentages applying, as for a biased threshold, and
with `log` nothing more is done. Whatever the action, the request's log line
names the lists, and each raises an alert.
- Checks the client's own address against the DNSBL zones `SWWAF_DNSBL_ZONES`
names, after the blocklists and before the rate limits (see "DNS blocklists"
below), by the verdicts it keeps. A zone is asked about a client in the
background, and no request waits for its answer: the request that has it
asked, and any other from the client before the answer comes, goes on as from
a client the zone does not list. With `SWWAF_REPUTATION_ACTION` at `limit:25`,
its default, a client a zone's verdict lists gets a quarter of every rate
limit and byte limit, the lowest of its percentages applying, as for a biased
threshold. With `deny`, its requests are refused with `SWWAF_BAN_RESPONSE`
before their bodies are read; they are not counted for the rate limits, and
make no ban. With `log`, nothing more is done. Whatever the action, the
request's log line names the zones, and each raises an alert. A client a
blocklist refuses is not checked.
- Checks the client's own address against the static lists, the three netblock
settings below, before anything else, its lookup included. A client in
`SWWAF_ALLOW_NETS` skips bans, the country lists, the blocklists, the rate
limits, the byte limits and the rule files, and is not looked up; the timeouts
and size limits still apply. A client in `SWWAF_DENY_NETS` is refused with
`SWWAF_BAN_RESPONSE` before its body is read, and the request is not counted
for the rate limits; an address in `SWWAF_ALLOW_NETS` too is let through. A
client in `SWWAF_RATE_LIMIT_EXEMPT_NETS` is neither counted nor refused by the
rate limits, and has no bytes counted by the byte limits; the country lists,
the rule files and bans still apply to it.
`SWWAF_ALLOW_NETS` skips bans, the country lists, the blocklists, the DNSBL
zones, the rate limits, the byte limits and the rule files, and is not looked
up; the timeouts and size limits still apply. A client in `SWWAF_DENY_NETS` is
refused with `SWWAF_BAN_RESPONSE` before its body is read, and the request is
not counted for the rate limits; an address in `SWWAF_ALLOW_NETS` too is let
through. A client in `SWWAF_RATE_LIMIT_EXEMPT_NETS` is neither counted nor
refused by the rate limits, and has no bytes counted by the byte limits; the
country lists, the rule files and bans still apply to it.
- In `observe` mode, with `SWWAF_MODE=observe`, refuses none of the requests
that `SWWAF_DENY_NETS`, a ban, the country lists, a blocklist, a rate limit or
a rule would refuse: it passes them to the app, and their log lines name what
`enforce` mode would have done (see `would_action` in "Request log" below).
The checks run, and requests and bytes are counted, as in `enforce` mode, with
three differences: neither a broken rate limit or byte limit nor a `ban` rule
makes a ban; a broken limit does not set the client's counters back to zero,
so each request over a rate limit is logged as one that would be refused, and
each whose bytes keep the client over a byte limit as breaking it; and a
request under a ban does not make it permanent. As in `enforce` mode, the
bytes counted are only those of the requests `enforce` mode would have passed
to the app. A ban it would have made, or made permanent, raises the alert
`enforce` mode would have raised, marked as what would have happened (see
"Alerts" below). The bans in `bans.json` are kept, and refuse requests again
when `smallwebwaf` next runs in `enforce` mode, as long as they last. The
timeouts and size limits still apply, since they protect `smallwebwaf` and the
app themselves, and a request for one of `smallwebwaf`'s own endpoints without
its token is still answered `401`. It is for trying a configuration before
enforcing it.
that `SWWAF_DENY_NETS`, a ban, the country lists, a blocklist, a DNSBL zone's
verdict, a rate limit or a rule would refuse: it passes them to the app, and
their log lines name what `enforce` mode would have done (see `would_action`
in "Request log" below). The checks run, and requests and bytes are counted,
as in `enforce` mode, with three differences: neither a broken rate limit or
byte limit nor a `ban` rule makes a ban; a broken limit does not set the
client's counters back to zero, so each request over a rate limit is logged as
one that would be refused, and each whose bytes keep the client over a byte
limit as breaking it; and a request under a ban does not make it permanent. As
in `enforce` mode, the bytes counted are only those of the requests `enforce`
mode would have passed to the app. A ban it would have made, or made
permanent, raises the alert `enforce` mode would have raised, marked as what
would have happened (see "Alerts" below). The bans in `bans.json` are kept,
and refuse requests again when `smallwebwaf` next runs in `enforce` mode, as
long as they last. The timeouts and size limits still apply, since they
protect `smallwebwaf` and the app themselves, and a request for one of
`smallwebwaf`'s own endpoints without its token is still answered `401`. It is
for trying a configuration before enforcing it.
- Answers `GET /_smallwebwaf/healthz` itself with `200` and `ok`, before any
check and without asking the app, for the image's health check.
- Answers `GET /_smallwebwaf/metrics` with its metrics (see "Metrics" below) for
@@ -254,14 +270,14 @@ in `bin/state` unless `SWWAF_STATE_DIR` is set, and the default rule file of
`SWWAF_LOG_REMOTE_URL` names one (see "Sending the log to a syslog server"
below).
- Sends an alert for each ban it makes or makes permanent, for a count over an
anomaly threshold, for a client a blocklist lists, for GeoJS failing or a list
it cannot fetch, for a rule file or state file with an error, and for a
replacement of the lookup database it cannot read, holding back repeats and,
past an hourly limit, rolling the rest into one summary, to each destination
you name: as a JSON object to the webhook `SWWAF_ALERT_WEBHOOK_URL` names, as
a message to the Slack incoming webhook `SWWAF_ALERT_SLACK_WEBHOOK_URL` names,
and as a message to the ntfy topic `SWWAF_ALERT_NTFY_URL` names (see "Alerts"
below).
anomaly threshold, for a client a blocklist or a DNSBL zone lists, for GeoJS
failing, a list it cannot fetch or a DNSBL zone that fails or refuses a query,
for a rule file or state file with an error, and for a replacement of the
lookup database it cannot read, holding back repeats and, past an hourly
limit, rolling the rest into one summary, to each destination you name: as a
JSON object to the webhook `SWWAF_ALERT_WEBHOOK_URL` names, as a message to
the Slack incoming webhook `SWWAF_ALERT_SLACK_WEBHOOK_URL` names, and as a
message to the ntfy topic `SWWAF_ALERT_NTFY_URL` names (see "Alerts" below).
- Counts requests and their bytes over a minute and an hour, per client, per
netblock around a client, per AS number, for the whole service and per named
netblock, and sends an `anomaly` alert for a count over the anomaly threshold
@@ -324,8 +340,8 @@ effective settings are logged at start.
- `SWWAF_REQUEST_MAX_BYTES` (default `100M`): the largest request body.
- `SWWAF_RESPONSE_MAX_BYTES` (default `5G`): the largest response body.
- `SWWAF_ALLOW_NETS` (default empty): netblocks whose clients skip bans, the
country lists, the blocklists, the rate limits, the byte limits and the rule
files, such as your monitoring or your own networks.
country lists, the blocklists, the DNSBL zones, the rate limits, the byte
limits and the rule files, such as your monitoring or your own networks.
- `SWWAF_RATE_LIMIT_EXEMPT_NETS` (default empty): netblocks whose clients the
rate limits and the byte limits do not apply to, such as a machine that talks
to the app all day.
@@ -427,10 +443,34 @@ effective settings are logged at start.
`limit:<percent>`, such as `limit:25`, gives it that percentage of every rate
limit and byte limit, and `log` does nothing more than note the lists in the
log line and raise the alert.
- `SWWAF_DNSBL_ZONES` (default empty): the DNSBL zones each client is asked
about, such as `dnsbl.dronebl.org` (see "DNS blocklists" below). A zone is a
DNS name of at most 189 characters: labels of letters, digits and hyphens, of
up to 63 characters each, neither starting nor ending with a hyphen, joined by
dots. A Spamhaus zone is the name of its keyed query service, with the key in
it, such as `<key>.xbl.dq.spamhaus.net`, and is shown with its key masked (see
"DNS blocklists" below). A zone that is not such a name stops the start, as
does one listed twice, even with its letters in another case or with another
key. Do not name a zone meant for mail (see "DNS blocklists" below).
- `SWWAF_DNSBL_RESOLVER` (default unset): the resolver the zones are asked
through, an IP address with an optional port, 53 when none is given, such as
`192.0.2.53` or `[2001:db8::53]:5353`. Unset, it is the host's, as
`/etc/resolv.conf` names it. Several zones refuse queries that come through a
public resolver.
- `SWWAF_REPUTATION_ACTION` (default `limit:25`): what is done with a client a
zone's verdict lists, as `SWWAF_BLOCKLIST_ACTION` is for a blocklist: `deny`,
`limit:<percent>` or `log`. A zone's verdict is less certain than a list such
as DROP, so by default a client it lists gets a quarter of every rate limit
and byte limit.
- `SWWAF_REPUTATION_CACHE_TTL` (default `24h`): how long a zone's verdict on a
client is used after the zone gave it.
- `SWWAF_REPUTATION_TIMEOUT` (default `2s`): how long a query to a zone may take
before it fails.
- `SWWAF_BAN_RESPONSE` (default `403`): how a refused client is answered, one
that is banned, breaks a rate limit, matches a `ban` rule, is in
`SWWAF_DENY_NETS`, comes from a refused country or is in a blocklist while
`SWWAF_BLOCKLIST_ACTION` is `deny`: `403`, `429`, or `close` to close the
`SWWAF_DENY_NETS`, comes from a refused country, is in a blocklist while
`SWWAF_BLOCKLIST_ACTION` is `deny` or is listed by a DNSBL zone while
`SWWAF_REPUTATION_ACTION` is `deny`: `403`, `429`, or `close` to close the
connection without an answer. Behind traefik, `close` does not leave the
client unanswered: traefik answers `502`, as it does whenever its backend
drops a connection. A `block` rule always answers `403`.
@@ -576,14 +616,16 @@ listed twice in one of them stops the start. `off` switches a timeout, a size
limit, a rate limit, a byte limit, an anomaly threshold, `SWWAF_ALERT_COOLDOWN`
or `SWWAF_ALERT_MAX_PER_HOUR` off; `SWWAF_CLIENT_REQUEST_HEADER_MAX_BYTES`,
`SWWAF_LOOKUP_TIMEOUT`, `SWWAF_UNKNOWN_LIMIT_PERCENT`,
`SWWAF_BLOCKLIST_REFRESH`, the ban settings, the state settings,
`SWWAF_BLOCKLIST_REFRESH`, `SWWAF_REPUTATION_CACHE_TTL`,
`SWWAF_REPUTATION_TIMEOUT`, the ban settings, the state settings,
`SWWAF_METRICS_TOP_N`, `SWWAF_LOG_REMOTE_BUFFER`, `SWWAF_ANOMALY_NET_V4_PREFIX`
and `SWWAF_ANOMALY_NET_V6_PREFIX` cannot be off.
Several limits are fixed rather than settings. At most 20,000 clients are kept,
with their counters and history, and an IPv6 client is counted by its /64. At
most 100,000 answers from GeoJS are kept, for 7 days each, and at most 20,000
anomaly counters.
most 100,000 answers from GeoJS are kept, for 7 days each, at most 20,000
anomaly counters, and at most 100,000 verdicts of the DNSBL zones, with at most
1,000 queries to them under way at once.
### Settings given as files
@@ -666,8 +708,9 @@ which every line has.
app's, as passed on, or those of `smallwebwaf`'s own answer.
- `request_bytes` and `response_bytes` count body bytes.
- `action` is `forward` for a request passed to the app, `denied` for one
refused because its client is in `SWWAF_DENY_NETS`, or in a blocklist while
`SWWAF_BLOCKLIST_ACTION` is `deny`, `banned` for one refused because a ban
refused because its client is in `SWWAF_DENY_NETS`, in a blocklist while
`SWWAF_BLOCKLIST_ACTION` is `deny`, or listed by a DNSBL zone while
`SWWAF_REPUTATION_ACTION` is `deny`, `banned` for one refused because a ban
covers its client or because it matched a `ban` rule, which bans its client,
`country_denied` for one refused for its client's country, `rate_limited` for
one that broke a rate limit and banned its client, `rule_blocked` for one a
@@ -676,16 +719,17 @@ which every line has.
could not be reached or its answer broke off, and `admin` for one
`smallwebwaf` answered at its own endpoint.
- `would_action` is there in `observe` mode for a request that
`SWWAF_DENY_NETS`, a ban, the country lists, a blocklist, a rate limit or a
rule would have refused in `enforce` mode, and names the action that refusal
would have had: `denied`, `banned`, `country_denied`, `rate_limited` or
`rule_blocked`. `action` then names what was done: `forward` for a request
passed to the app, and another action, such as `too_large`, for one a size or
time limit refused.
`SWWAF_DENY_NETS`, a ban, the country lists, a blocklist, a DNSBL zone's
verdict, a rate limit or a rule would have refused in `enforce` mode, and
names the action that refusal would have had: `denied`, `banned`,
`country_denied`, `rate_limited` or `rule_blocked`. `action` then names what
was done: `forward` for a request passed to the app, and another action, such
as `too_large`, for one a size or time limit refused.
- `limit_percent` is there for a request the rate limits count whose client a
biased threshold, or `SWWAF_BLOCKLIST_ACTION` for a blocklist that lists it,
gives less than the whole of the rate limits, and gives the percentage it
gets, with `limit_percent_setting` naming the setting that gave it, such as
biased threshold, `SWWAF_BLOCKLIST_ACTION` for a blocklist that lists it, or
`SWWAF_REPUTATION_ACTION` for a DNSBL zone whose verdict lists it, gives less
than the whole of the rate limits, and gives the percentage it gets, with
`limit_percent_setting` naming the setting that gave it, such as
`SWWAF_ASN_LIMIT_PERCENT`, or `SWWAF_ASN_LIMIT_PERCENT_URL` for the file it
names. `bytes_percent` and `bytes_percent_setting` are the same for the byte
limits. Each is left out when the client gets the whole of those limits.
@@ -697,11 +741,11 @@ which every line has.
health check, one from a client in `SWWAF_ALLOW_NETS` or
`SWWAF_RATE_LIMIT_EXEMPT_NETS`, one for a path that
`SWWAF_RATE_LIMIT_EXEMPT_PATHS` exempts, and one that `SWWAF_DENY_NETS`, a
ban, the country lists or a blocklist refuse, or would refuse in `observe`
mode. Its `minute_bytes`, `hour_bytes` and `day_bytes` give the client's bytes
in each window as the byte limits count them, in the same way: for a request
whose bytes they count, with its own, once it has ended; for any other, those
counted before it.
ban, the country lists, a blocklist or a DNSBL zone's verdict refuse, or would
refuse in `observe` mode. Its `minute_bytes`, `hour_bytes` and `day_bytes`
give the client's bytes in each window as the byte limits count them, in the
same way: for a request whose bytes they count, with its own, once it has
ended; for any other, those counted before it.
- `rule_ids` is there for a request that matched rules of the rule files, and
lists their ids in the order they matched, up to the one that refused it.
- `limit_hit` is there for a request that broke a rate limit, or whose bytes
@@ -711,12 +755,16 @@ which every line has.
several. `offence` is then `limit`. A request whose bytes broke a byte limit
is not refused: its `action` is what it would have been otherwise, such as
`forward`.
- `reputation` is there for a request whose client a blocklist lists, and gives
the URLs of the blocklists that list it, in the order `SWWAF_BLOCKLIST_URLS`
names them, whatever `SWWAF_BLOCKLIST_ACTION` says. It is left out for a
client the blocklists are not checked for: one in `SWWAF_ALLOW_NETS`, and one
`SWWAF_DENY_NETS`, a ban or the country lists refuse first, or would in
`observe` mode.
- `reputation` is there for a request whose client a blocklist or a DNSBL zone's
verdict lists, and gives the URLs of the blocklists that list it, in the order
`SWWAF_BLOCKLIST_URLS` names them, then the zones whose verdict lists it, in
the order `SWWAF_DNSBL_ZONES` names them, whatever `SWWAF_BLOCKLIST_ACTION`
and `SWWAF_REPUTATION_ACTION` say. A zone whose verdict on the client has not
come yet, or was given `SWWAF_REPUTATION_CACHE_TTL` ago or more, is not named.
It is left out for a client the blocklists are not checked for: one in
`SWWAF_ALLOW_NETS`, and one `SWWAF_DENY_NETS`, a ban or the country lists
refuse first, or would in `observe` mode. The zones are not checked either for
a client a blocklist refuses, or would in `observe` mode.
- `ban_expires` is there for a request that made a ban or was refused under one,
or in `observe` mode would have been refused under one, and gives when the ban
ends, in the same form as `time`, or `permanent`.
@@ -786,11 +834,13 @@ it, as below. An alert is for one of these events, and is sent when
- `anomaly`: a count of requests or bytes over an anomaly threshold, raised by
each request that ends with the count over it, in `observe` mode as in
`enforce` mode. It refuses and bans nothing.
- `reputation_hit`: a request whose client a blocklist lists, one alert for each
blocklist that lists it, whatever `SWWAF_BLOCKLIST_ACTION` says, in `observe`
mode as in `enforce` mode.
- `source_failure`: GeoJS failing or refusing `smallwebwaf`, or a fetch of a
list failing (see "Blocklists" below).
- `reputation_hit`: a request whose client a blocklist or a DNSBL zone's verdict
lists, one alert for each blocklist and each zone that lists it, whatever
`SWWAF_BLOCKLIST_ACTION` and `SWWAF_REPUTATION_ACTION` say, in `observe` mode
as in `enforce` mode.
- `source_failure`: GeoJS failing or refusing `smallwebwaf`, a fetch of a list
failing (see "Blocklists" below), or a query to a DNSBL zone failing or
refused (see "DNS blocklists" below).
- `file_error`: a rule file edited while it runs that has an error, an edit of a
state file set aside as `<name>.bad`, a state file it could not write while
running, or a replacement of the lookup database it could not read, which it
@@ -863,7 +913,7 @@ is sent on one line:
- `reason` is a short sentence; for a ban, the ban's `reason` in `bans.json`;
for an `anomaly`, what was counted over which threshold, such as
`requests per minute of the netblock 203.0.113.0/24 over the threshold of 1000`;
for a `reputation_hit`, `listed by a blocklist`.
for a `reputation_hit`, `listed by a blocklist` or `listed by a DNSBL zone`.
- `detail` is what is particular to the event: for a ban, its `cause`, when it
ends as `ban_expires`, in the form the request log gives it, and its `notes`,
as `bans.json` gives them; for an `anomaly`, the `scope`, `client`, `net`,
@@ -871,11 +921,11 @@ is sent on one line:
`asn` and the `name` of the named netblock for `watch`, the `window`, `minute`
or `hour`, the `kind`, `requests` or `bytes`, the `count`, which is weighted
as the rate limits weigh theirs, and the `threshold`; for a `reputation_hit`,
the `source`, the URL of the blocklist; for `source_failure`, the `source`,
`geojs` or the URL of the list, the `error`, and for GeoJS, when it is asked
again, `asking_again_in`; for `file_error`, the `file`, which for an edit set
aside is the file it was renamed to, and the `error`, which for a file that
does not parse names where in it the error is.
the `source`, the URL of the blocklist or the zone; for `source_failure`, the
`source`, `geojs`, the URL of the list or the zone, the `error`, and for
GeoJS, when it is asked again, `asking_again_in`; for `file_error`, the
`file`, which for an edit set aside is the file it was renamed to, and the
`error`, which for a file that does not parse names where in it the error is.
- `suppressed_repeats` is how many repeats the cooldown held back before this
alert, and for a `summary`, those no other alert gives (see below).
@@ -921,11 +971,11 @@ and Slack this JSON object, shown indented; it is sent on one line:
```
An alert for the same event as the last one sent, on the same netblock, and for
a `reputation_hit` about the same blocklist, or for a `file_error` about the
same file, or for a `source_failure` about the same source, or for an `anomaly`
in the same scope, with the same netblock, AS number or name, whatever its
window and kind, less than `SWWAF_ALERT_COOLDOWN` after it, is a repeat: it is
held back and counted, and the next alert sent for them gives that count as
a `reputation_hit` about the same blocklist or zone, or for a `file_error` about
the same file, or for a `source_failure` about the same source, or for an
`anomaly` in the same scope, with the same netblock, AS number or name, whatever
its window and kind, less than `SWWAF_ALERT_COOLDOWN` after it, is a repeat: it
is held back and counted, and the next alert sent for them gives that count as
`suppressed_repeats`. As each hour of the clock, in UTC, ends, the cooldowns
that have run out are dropped, and the repeats they held back, which no alert
sent since has given, go in that hour's summary.
@@ -990,8 +1040,11 @@ with times in UTC.
indented to be read, under `lists`: its `url`, when it was last `tried`, the
fetch failed or not, and its last good copy: when that was `fetched`, and its
`lines`, as fetched, comment lines included, each on a line of its own, both
left out while no fetch of it has succeeded. As the file is read, the lists
the settings no longer name are dropped.
left out while no fetch of it has succeeded; and under `verdicts`, each
verdict of a DNSBL zone still in use (see "DNS blocklists" below): its `zone`,
the `client`'s address, whether the zone `listed` the client, and when the
zone gave it, `fetched`. As the file is read, the lists the settings no longer
name, and the verdicts of the zones they no longer name, are dropped.
- `alerts.json`: the state of the alerts (see "Alerts" above), indented to be
read: under `cooldowns`, for each event and netblock, with the `source` too
for a `reputation_hit`, or event and `file` or `source`, or for an `anomaly`,
@@ -1029,24 +1082,26 @@ keeps refusing every client in its netblock until it ends, even after
until a fetch of it succeeds. A netblock whose address has bits past its length,
such as `203.0.113.9/24`, is read as the netblock it is in, `203.0.113.0/24`.
Buckets and answers whose time has passed are dropped, and so is an anomaly
counter left with no bucket. A missing file is empty state, as on a first start.
A file that does not parse, or has another `version`, stops the start with a
message naming the file, and the line and column where Go's JSON decoder gives
them; so does a state directory `smallwebwaf` cannot write. So does an entry
without a field it needs, named with the entry's place in the file: a ban's
`netblock`, `start` or `expires`, which is `null` for a permanent ban; a
client's `client`, or the `start` of a window in which it has requests or bytes;
an answer's `client`, `country`, which is `""` for a client GeoJS cannot place,
or `answered`; a cooldown's `event` or `sent`; an alert waiting's `event` or
`time`; an anomaly counter's `netblock`, unless it counts an AS number or the
whole service, its `asn`, for an AS number, its `name`, for a named netblock, or
the `start` of a window in which it has requests or bytes; a list's `url`,
`fetched` or `lines`, which is `[]` for an empty list. So does a ban whose
`cause` is not `limit`, `attack` or `admin`, alerts waiting for a destination
that is not `webhook`, `slack` or `ntfy`, an anomaly counter whose `scope` is
not `client`, `net`, `asn`, `total` or `watch`, and a copy of a list with a line
that would make its fetch fail. An answer's `asn` or `as_name` left out reads as
empty.
counter left with no bucket; a verdict past its time is neither used nor written
again. A missing file is empty state, as on a first start. A file that does not
parse, or has another `version`, stops the start with a message naming the file,
and the line and column where Go's JSON decoder gives them; so does a state
directory `smallwebwaf` cannot write. So does an entry without a field it needs,
named with the entry's place in the file: a ban's `netblock`, `start` or
`expires`, which is `null` for a permanent ban; a client's `client`, or the
`start` of a window in which it has requests or bytes; an answer's `client`,
`country`, which is `""` for a client GeoJS cannot place, or `answered`; a
cooldown's `event` or `sent`; an alert waiting's `event` or `time`; an anomaly
counter's `netblock`, unless it counts an AS number or the whole service, its
`asn`, for an AS number, its `name`, for a named netblock, or the `start` of a
window in which it has requests or bytes; a list's `url`, `fetched` or `lines`,
which is `[]` for an empty list; a verdict's `zone`, `client`, `listed`, which
is `false` for a client the zone does not list, or `fetched`. So does a ban
whose `cause` is not `limit`, `attack` or `admin`, alerts waiting for a
destination that is not `webhook`, `slack` or `ntfy`, an anomaly counter whose
`scope` is not `client`, `net`, `asn`, `total` or `watch`, and a copy of a list
with a line that would make its fetch fail. An answer's `asn` or `as_name` left
out reads as empty.
While it runs, `smallwebwaf` watches `SWWAF_STATE_DIR` and takes in your edit of
a state file as soon as you save it: what the file then holds replaces what
@@ -1238,6 +1293,11 @@ scraped, and keeps this one as `exported_instance` unless the scrape sets
`smallwebwaf_reputation_failures_total`: the fetches of the list that failed;
and `smallwebwaf_reputation_last_fetch_timestamp_seconds`: when the copy of it
in use was fetched, `0` while there is none.
- By `source`, each zone `SWWAF_DNSBL_ZONES` names:
`smallwebwaf_reputation_hits_total`: the requests whose client the zone's
verdict lists, a series that comes with the first;
`smallwebwaf_reputation_queries_total`: the queries made to the zone; and
`smallwebwaf_reputation_failures_total`: those that failed.
- `smallwebwaf_tracked_clients`: the clients in the table of clients.
- `smallwebwaf_state_file_writes_total`,
`smallwebwaf_state_file_write_failures_total`,
@@ -1429,9 +1489,9 @@ goes through the candidates one by one.
readable JSON files, written regularly and at every stop, so a restart loses
nothing. Edit a file, or add a rule file, and the running `smallwebwaf` picks
up the change. Nothing is read from disk while serving a request. The files
for the bans, the clients, the GeoJS answers, the copies of the lists and the
alerts are built, with an edit taken in while running (see "State files"
above); the rest comes with its features.
for the bans, the clients, the GeoJS answers, the copies of the lists, the
verdicts of the DNSBL zones and the alerts are built, with an edit taken in
while running (see "State files" above); the rest comes with its features.
- Health checks, the metrics, and listing, adding and lifting bans or asking why
a given address was refused, all on the one port every request uses: under
`/_smallwebwaf/` on the app's own address, through traefik like any other
@@ -1659,6 +1719,58 @@ hour apart whatever `SWWAF_BLOCKLIST_REFRESH` is: each fetches a list again that
long after its own last try, so those first started within the same hour, with
the same refresh, keep fetching within the same hour.
## DNS blocklists
`SWWAF_DNSBL_ZONES` names DNS blocklists, DNSBL zones, which list an address by
answering a query for a name made from it. None is named by default, for the
reason none of the blocklists is. `smallwebwaf` asks each zone about a client's
own address in the background, the first time it sees the client: the request
goes on at once, as from a client the zone does not list, and so does every
request from it until the zone has answered. The name asked about is the one RFC
5782 gives: the four numbers of an IPv4 address in reverse order, so that
`192.0.2.99` is asked about in `dnsbl.dronebl.org` as
`99.2.0.192.dnsbl.dronebl.org`, or the 32 hex digits of an IPv6 address in
reverse order, each followed by a dot. An IPv6 client is asked about by its own
address, not by its /64.
A zone that answers that the name does not exist, or has no address, does not
list the client, and one that answers with an address in `127.0.0.0/8` lists it.
Any other answer gives no verdict, and is a failure: an address in
`127.255.255.0/24`, with which Spamhaus refuses a query, such as one sent
through a public resolver or one past its limit; an address outside
`127.0.0.0/8`, such as a resolver gives that answers even for names that do not
exist; an error the resolver answers with, such as a refusal; and no answer
within `SWWAF_REPUTATION_TIMEOUT`. A failure is counted, logged and raised as a
`source_failure` alert, held back as a repeat within `SWWAF_ALERT_COOLDOWN`, and
the zone is not asked again for a minute, so that a zone that refuses queries is
not asked on every request.
Each verdict, the zone listing the client or not, is kept, in memory and in
`reputation.json` (see "State files" above), so that a restart keeps it, and is
used for `SWWAF_REPUTATION_CACHE_TTL` after the zone gave it, 24 hours by
default. The client's first request after that has the zone asked again, and
until it answers, no verdict of the zone applies to the client. At most 100,000
verdicts are kept, the one fetched longest ago dropped first, and at most 1,000
queries are under way at once; past that, a zone is asked about a client at the
client's next request.
A client in `SWWAF_ALLOW_NETS` is not checked, nor is one a blocklist refuses.
`SWWAF_REPUTATION_ACTION` says what is done with one a zone's verdict lists, as
"What it does so far" above describes.
Do not name a zone meant for mail, such as one of residential and dynamic
address ranges, which ordinary visitors come from, or one that includes such a
list, as Spamhaus's `zen` does: it would refuse them, or lower their limits.
Spamhaus's zones answer through its keyed query service, named with the key in
it, such as `<key>.xbl.dq.spamhaus.net`. The key of a zone under
`dq.spamhaus.net` is its first label, and `smallwebwaf` shows `********` in its
place wherever it names the zone, as `********.xbl.dq.spamhaus.net`: in the
settings logged at start, an error that stops the start, its own messages, the
request log, the alerts and the metrics. Only `reputation.json` keeps the zone
with its key. A key in the name of any other zone is shown as given. Several
zones refuse queries that come through a public resolver; `SWWAF_DNSBL_RESOLVER`
names another resolver to ask through.
## How the code is laid out
- `cmd/smallwebwaf`: the binary, which only calls `internal/smallwebwaf`.
@@ -1672,15 +1784,15 @@ the same refresh, keep fetching within the same hour.
standard library's `httputil.ReverseProxy` within the timeouts and size
limits, and writes the request's log line. Its `check` method is where a
request is refused before anything reaches the app: for `SWWAF_DENY_NETS`, for
a ban, for the country lists, for a blocklist, for a rate limit, which bans
the client, for a `block` or `ban` rule, the latter banning the client, and
for an announced body over the size limit; in `observe` mode, only for the
size limit, with what it would have refused for noted in the log line. A
request under `/_smallwebwaf/` that `check` lets through is answered by
`answerAdmin` instead of reaching the app. Once the answer to a request passed
to the app has ended, `countBytes` counts its bytes for the byte limits, and
once any request but the health check has ended, `countAnomalies` counts it
for the anomaly thresholds.
a ban, for the country lists, for a blocklist, for a DNSBL zone's verdict, for
a rate limit, which bans the client, for a `block` or `ban` rule, the latter
banning the client, and for an announced body over the size limit; in
`observe` mode, only for the size limit, with what it would have refused for
noted in the log line. A request under `/_smallwebwaf/` that `check` lets
through is answered by `answerAdmin` instead of reaching the app. Once the
answer to a request passed to the app has ended, `countBytes` counts its bytes
for the byte limits, and once any request but the health check has ended,
`countAnomalies` counts it for the anomaly thresholds.
- `internal/metrics`: the metrics, counted as the other parts tell it what
happened, and served in the Prometheus text format.
- `internal/bans`: the ban ledger: each netblock's bans with their notes, how
@@ -1696,7 +1808,9 @@ the same refresh, keep fetching within the same hour.
- `internal/reputation`: fetches the blocklists and the file
`SWWAF_ASN_LIMIT_PERCENT_URL` names as they are due, keeps the last good copy
of each, and tells which blocklists list an address and what percentage the
file gives an AS number.
file gives an AS number; and asks the DNSBL zones about clients in the
background, through the standard library's resolver, keeps their verdicts, and
tells which zones' verdicts list an address.
- `internal/ratelimit`: the table of clients: counts each client's requests and
bytes, tells when they take it over a rate limit or a byte limit, and keeps
each client's history.
@@ -1724,7 +1838,8 @@ the same refresh, keep fetching within the same hour.
Besides the Go standard library, `github.com/hashicorp/golang-lru/v2` keeps the
table of clients to 20,000 and the GeoJS answers to 100,000, dropping the least
recently seen, the anomaly counters to 20,000, dropping the one counted least
recently seen, the DNSBL zones' verdicts to 100,000, dropping the one fetched
longest ago, the anomaly counters to 20,000, dropping the one counted least
recently, and the banned netblocks in the order they were last seen, from which
the ledger picks the ban to drop past `SWWAF_MAX_BANS`, and
`github.com/prometheus/client_golang` keeps the metrics and serves them, and
internal/alerts/alerts.go
+4 -3
View File
@@ -46,10 +46,11 @@ const (
EventAnomaly = "anomaly"
// EventWAFBlock comes with the Core Rule Set; nothing raises it yet.
EventWAFBlock = "waf_block"
// EventReputationHit is a request whose client a blocklist lists.
// EventReputationHit is a request whose client a blocklist or a DNSBL
// zone lists.
EventReputationHit = "reputation_hit"
// EventSourceFailure is GeoJS failing or refusing smallwebwaf, or a
// fetch of a list failing.
// EventSourceFailure is GeoJS failing or refusing smallwebwaf, a fetch
// of a list failing, or a query to a DNSBL zone failing or refused.
EventSourceFailure = "source_failure"
// EventFileError is a rule file or state file edited while smallwebwaf
// runs that does not parse, a replacement of the lookup database that
internal/config/config.go
+163 -3
View File
@@ -152,6 +152,21 @@ type Config struct {
BlocklistRefresh time.Duration
BlocklistAction string
BlocklistLimitPercent int64
// DNSBLZones are the DNSBL zones clients are asked about
// (SWWAF_DNSBL_ZONES), through DNSBLResolver (SWWAF_DNSBL_RESOLVER), or
// the host's resolver while that is the zero AddrPort.
// ReputationAction is what is done with a client a zone's verdict lists
// (SWWAF_REPUTATION_ACTION): deny, limit or log; for limit,
// ReputationLimitPercent is the percentage of every limit it gets. A
// verdict is used for ReputationCacheTTL after it was fetched
// (SWWAF_REPUTATION_CACHE_TTL), and a query may take ReputationTimeout
// (SWWAF_REPUTATION_TIMEOUT). Neither can be off.
DNSBLZones []string
DNSBLResolver netip.AddrPort
ReputationAction string
ReputationLimitPercent int64
ReputationCacheTTL time.Duration
ReputationTimeout time.Duration
// BanResponse is the status a refused client is answered with, 403
// or 429, or 0 to close the connection without an answer
// (SWWAF_BAN_RESPONSE). It answers a banned client, a request that
@@ -366,6 +381,11 @@ var (
errNotAnHourOrMore = errors.New("is not a duration of 1h or more, such as 24h")
errNotAction = errors.New(
"is not deny, limit:<percent> such as limit:25, or log")
errNotZone = errors.New("is not a DNS zone such as dnsbl.dronebl.org")
errZoneTooLong = errors.New("is longer than 189 characters, too long for the " +
"names IPv6 clients are asked about by")
errNotResolver = errors.New("is not an IP address with an optional port, " +
"such as 192.0.2.53 or [2001:db8::53]:5353")
)
// FromEnvironment reads the settings with lookupEnv, normally
@@ -418,6 +438,10 @@ func FromEnvironment(lookupEnv func(string) (string, bool)) (*Config, error) {
ASNLimitPercentURL: env.listURL("SWWAF_ASN_LIMIT_PERCENT_URL"),
BlocklistURLs: env.listURLs("SWWAF_BLOCKLIST_URLS"),
BlocklistRefresh: env.refresh("SWWAF_BLOCKLIST_REFRESH", "24h"),
DNSBLZones: env.zones("SWWAF_DNSBL_ZONES"),
DNSBLResolver: env.resolver("SWWAF_DNSBL_RESOLVER"),
ReputationCacheTTL: env.durationNotOff("SWWAF_REPUTATION_CACHE_TTL", "24h"),
ReputationTimeout: env.durationNotOff("SWWAF_REPUTATION_TIMEOUT", "2s"),
BanResponse: env.banResponse("SWWAF_BAN_RESPONSE", "403"),
LimitBanDuration: env.durationNotOff("SWWAF_LIMIT_BAN_DURATION", "1h"),
LimitBanRepeatWindow: env.durationNotOff("SWWAF_LIMIT_BAN_REPEAT_WINDOW", "24h"),
@@ -462,6 +486,8 @@ func FromEnvironment(lookupEnv func(string) (string, bool)) (*Config, error) {
cfg.InstanceName, cfg.LogRemoteURL != nil)
cfg.BlocklistAction, cfg.BlocklistLimitPercent = env.action(
"SWWAF_BLOCKLIST_ACTION", "deny")
cfg.ReputationAction, cfg.ReputationLimitPercent = env.action(
"SWWAF_REPUTATION_ACTION", "limit:25")
env.checkInstanceNameForNtfy(cfg.InstanceName, cfg.AlertNtfyURL != nil)
env.checkLookupDBPath(cfg)
@@ -733,9 +759,9 @@ func (e *environment) refresh(name, defaultValue string) time.Duration {
return duration
}
// action reads a setting that is what is done with a client a list names:
// deny, log, or limit:<percent>, which it returns as limit and the
// percentage.
// action reads a setting that is what is done with a client a blocklist
// or a DNSBL zone lists: deny, log, or limit:<percent>, which it returns
// as limit and the percentage.
func (e *environment) action(name, defaultValue string) (string, int64) {
value := e.value(name, defaultValue)
if value == "deny" || value == "log" {
@@ -752,6 +778,33 @@ func (e *environment) action(name, defaultValue string) (string, int64) {
return "limit", percent
}
// zones reads the setting that is the list of DNSBL zones. It is empty by
// default. The log shows each zone with its key masked, as MaskZoneKey
// masks it.
func (e *environment) zones(name string) []string {
value, _ := e.lookup(name)
zones, err := parseZones(value)
e.check(name, err)
logged := make([]string, len(zones))
for i, zone := range zones {
logged[i] = MaskZoneKey(zone)
}
e.settings = append(e.settings, slog.String(name, strings.Join(logged, ",")))
return zones
}
// resolver reads the setting that is the resolver the DNSBL zones are
// asked through, the zero AddrPort while it is unset or empty.
func (e *environment) resolver(name string) netip.AddrPort {
resolver, err := parseResolver(e.value(name, ""))
e.check(name, err)
return resolver
}
// lookupSource reads the setting that is where clients are looked up:
// geojs, file, or off.
func (e *environment) lookupSource(name, defaultValue string) string {
@@ -1697,6 +1750,113 @@ func parseListURLs(value string) ([]string, error) {
return urls, nil
}
const (
// maxZoneLength is the most characters a DNSBL zone may have: 253, the
// most a DNS name may have, less the 64 that come before the zone in
// the name an IPv6 client is asked about by, its 32 hex digits each
// followed by a dot.
maxZoneLength = 189
// maxLabelLength is the most characters a label of a DNS name may have.
maxLabelLength = 63
// labelChars are the characters a label of a DNS zone may hold.
labelChars = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789-"
// dnsPort is the port a resolver is asked on when SWWAF_DNSBL_RESOLVER
// gives none.
dnsPort = 53
)
// parseZones reads a comma-separated list of DNSBL zones, each a DNS name
// such as dnsbl.dronebl.org: labels separated by dots, each of 1 to 63
// letters, digits and hyphens, neither starting nor ending with a hyphen,
// and at most maxZoneLength characters in all. Go's resolver takes any
// other name for one that does not exist, so that the zone would list no
// client. A zone listed twice is an error, whatever the case of its
// letters, which DNS names ignore, and whatever its key, since
// MaskZoneKey shows two keys of one zone alike. An error shows a zone as
// MaskZoneKey does.
func parseZones(value string) ([]string, error) {
zones, err := parseList(value)
if err != nil {
// parseList's error, for an empty item, shows the whole value, keys
// included.
return nil, errEmptyItem
}
for i, zone := range zones {
shown := MaskZoneKey(zone)
listedBefore := slices.ContainsFunc(zones[:i], func(earlier string) bool {
return strings.EqualFold(MaskZoneKey(earlier), shown)
})
switch {
case len(zone) > maxZoneLength:
return nil, fmt.Errorf("%q %w", shown, errZoneTooLong)
case !isZone(zone):
return nil, fmt.Errorf("%q %w", shown, errNotZone)
case listedBefore:
return nil, fmt.Errorf("%q %w", shown, errListedTwice)
}
}
return zones, nil
}
// MaskZoneKey returns zone with ******** in place of its key, if it is a
// zone of Spamhaus's keyed query service, a name under dq.spamhaus.net,
// such as <key>.xbl.dq.spamhaus.net, whose first label is the key. Any
// other zone it returns as it is. A zone is shown so wherever it leaves
// the process: in the log, the alerts and the metrics.
func MaskZoneKey(zone string) string {
// DNS names ignore case, and a name may be written with a dot at its
// end.
name := strings.TrimSuffix(strings.ToLower(zone), ".")
if !strings.HasSuffix(name, ".dq.spamhaus.net") {
return zone
}
_, rest, _ := strings.Cut(zone, ".")
return masked + "." + rest
}
// isZone reports whether each label of zone is as parseZones takes it.
func isZone(zone string) bool {
for label := range strings.SplitSeq(zone, ".") {
badChar := strings.ContainsFunc(label, func(char rune) bool {
return !strings.ContainsRune(labelChars, char)
})
if label == "" || len(label) > maxLabelLength || badChar ||
strings.HasPrefix(label, "-") || strings.HasSuffix(label, "-") {
return false
}
}
return true
}
// parseResolver reads the resolver the DNSBL zones are asked through: an
// IP address with a port from 1 to 65535, such as 192.0.2.53:5353 or
// [2001:db8::53]:5353, or without one, such as 192.0.2.53 or 2001:db8::53,
// for port 53. An empty value is none, the zero AddrPort.
func parseResolver(value string) (netip.AddrPort, error) {
if value == "" {
return netip.AddrPort{}, nil
}
resolver, err := netip.ParseAddrPort(value)
if err == nil && resolver.Port() != 0 {
return resolver, nil
}
addr, err := netip.ParseAddr(value)
if err != nil {
return netip.AddrPort{}, fmt.Errorf("%q %w", value, errNotResolver)
}
return netip.AddrPortFrom(addr, dnsPort), nil
}
// parseWebhookHeaders reads a comma-separated list of headers, each its
// name, :, and its value, and returns them, and how the log shows them,
// with each value as ********. An error names the item by its place in
internal/config/config_test.go
+210
View File
@@ -61,6 +61,11 @@ const (
blocklistURLs = "SWWAF_BLOCKLIST_URLS"
blocklistRefresh = "SWWAF_BLOCKLIST_REFRESH"
blocklistAction = "SWWAF_BLOCKLIST_ACTION"
dnsblZones = "SWWAF_DNSBL_ZONES"
dnsblResolver = "SWWAF_DNSBL_RESOLVER"
reputationAction = "SWWAF_REPUTATION_ACTION"
reputationCacheTTL = "SWWAF_REPUTATION_CACHE_TTL"
reputationTimeout = "SWWAF_REPUTATION_TIMEOUT"
banResponse = "SWWAF_BAN_RESPONSE"
limitBanDuration = "SWWAF_LIMIT_BAN_DURATION"
limitBanRepeatWindow = "SWWAF_LIMIT_BAN_REPEAT_WINDOW"
@@ -152,6 +157,9 @@ const (
defaultAlertCooldown = "15m"
)
// defaultReputationCacheTTL is the default of SWWAF_REPUTATION_CACHE_TTL.
const defaultReputationCacheTTL = "24h"
// defaultLogRequestHeaders is the default of SWWAF_LOG_REQUEST_HEADERS.
const defaultLogRequestHeaders = "accept,accept-language,accept-encoding," +
"content-type,origin,range"
@@ -1322,6 +1330,202 @@ func TestASNLimitPercentURLThatIsABlocklistStopsTheStart(t *testing.T) {
}
}
// dronebl is a DNSBL zone, and spamhaus one of Spamhaus's, a name
// containing spamhausKey, the key of its keyed query service, which the
// log shows as spamhausMasked.
const (
dronebl = "dnsbl.dronebl.org"
spamhausKey = "abcdefghijklmnopqrstuvwxyz"
spamhaus = spamhausKey + ".xbl.dq.spamhaus.net"
spamhausMasked = "********.xbl.dq.spamhaus.net"
)
func TestDNSBLSettingsAsSet(t *testing.T) {
t.Parallel()
cfg := fromEnvironment(t, environment{})
if len(cfg.DNSBLZones) != 0 || cfg.DNSBLResolver.IsValid() ||
cfg.ReputationAction != actionLimit || cfg.ReputationLimitPercent != 25 ||
cfg.ReputationCacheTTL != 24*time.Hour || cfg.ReputationTimeout != 2*time.Second {
t.Errorf("by default, the zones %v, the resolver %s, the action %s:%d, the TTL %s "+
"and the timeout %s, want no zone, no resolver, limit:25, 24h and 2s",
cfg.DNSBLZones,
cfg.DNSBLResolver, cfg.ReputationAction, cfg.ReputationLimitPercent,
cfg.ReputationCacheTTL, cfg.ReputationTimeout)
}
// The longest zone allowed, of 189 characters, with labels of 63, the
// longest allowed.
longest := strings.Repeat("a", 63) + "." + strings.Repeat("b", 63) + "." +
strings.Repeat("c", 61)
for _, tc := range []struct {
zones, resolver, action string
// want are the zones, resolver, action and percent Config gives.
want []string
wantResolver string
wantAction string
wantPercent int64
}{
{
dronebl + ", " + spamhaus, "192.0.2.53", actionDeny,
[]string{dronebl, spamhaus}, "192.0.2.53:53", actionDeny, 0,
},
{
longest, "192.0.2.53:5353", actionLog,
[]string{longest}, "192.0.2.53:5353", actionLog, 0,
},
{
"Dnsbl-1.Example", "2001:db8::53", "limit:10",
[]string{"Dnsbl-1.Example"}, "[2001:db8::53]:53", actionLimit, 10,
},
{
dronebl, "[2001:db8::53]:5353", "limit:0",
[]string{dronebl}, "[2001:db8::53]:5353", actionLimit, 0,
},
} {
cfg := fromEnvironment(t, environment{
dnsblZones: tc.zones, dnsblResolver: tc.resolver, reputationAction: tc.action,
reputationCacheTTL: "12h", reputationTimeout: "3s",
})
if !slices.Equal(cfg.DNSBLZones, tc.want) ||
cfg.DNSBLResolver.String() != tc.wantResolver ||
cfg.ReputationAction != tc.wantAction ||
cfg.ReputationLimitPercent != tc.wantPercent ||
cfg.ReputationCacheTTL != 12*time.Hour ||
cfg.ReputationTimeout != 3*time.Second {
t.Errorf("%s=%s, %s=%s and %s=%s gave %v, %s, %s:%d, %s and %s", dnsblZones,
tc.zones, dnsblResolver, tc.resolver, reputationAction, tc.action,
cfg.DNSBLZones, cfg.DNSBLResolver, cfg.ReputationAction,
cfg.ReputationLimitPercent, cfg.ReputationCacheTTL, cfg.ReputationTimeout)
}
}
}
func TestInvalidDNSBLSettingStopsTheStartSayingWhatIsWrong(t *testing.T) {
t.Parallel()
const (
notZone = " is not a DNS zone such as dnsbl.dronebl.org"
notResolver = " is not an IP address with an optional port, such as 192.0.2.53 " +
"or [2001:db8::53]:5353"
notAboveZero = " is not a duration above zero, such as 1h or 7d"
)
label64 := strings.Repeat("a", 64) + ".example"
tooLong := strings.Repeat("a", 63) + "." + strings.Repeat("b", 63) + "." +
strings.Repeat("c", 62)
for _, tc := range []struct{ name, value, want string }{
{dnsblZones, "dnsbl..example", `"dnsbl..example"` + notZone},
{dnsblZones, "dnsbl.example.", `"dnsbl.example."` + notZone},
{dnsblZones, "-dnsbl.example", `"-dnsbl.example"` + notZone},
{dnsblZones, "dnsbl-.example", `"dnsbl-.example"` + notZone},
{dnsblZones, "dns_bl.example", `"dns_bl.example"` + notZone},
{dnsblZones, label64, `"` + label64 + `"` + notZone},
{
dnsblZones, tooLong,
`"` + tooLong + `" is longer than 189 characters, too long for the names ` +
`IPv6 clients are asked about by`,
},
{
dnsblZones, dronebl + "," + spamhaus + "," + dronebl,
`"` + dronebl + `" is listed twice`,
},
// DNS names ignore case.
{dnsblZones, "dnsbl.example,DNSBL.example", `"DNSBL.example" is listed twice`},
{dnsblResolver, "resolver.example", `"resolver.example"` + notResolver},
{dnsblResolver, "192.0.2.53:0", `"192.0.2.53:0"` + notResolver},
{dnsblResolver, "192.0.2.53:65536", `"192.0.2.53:65536"` + notResolver},
{dnsblResolver, "[2001:db8::53]", `"[2001:db8::53]"` + notResolver},
{
reputationAction, "ban",
`"ban" is not deny, limit:<percent> such as limit:25, or log`,
},
{reputationCacheTTL, off, `"off"` + notAboveZero},
{reputationTimeout, "0s", `"0s"` + notAboveZero},
} {
t.Run(tc.name+"="+tc.value, func(t *testing.T) {
t.Parallel()
_, err := config.FromEnvironment(environment{tc.name: tc.value}.lookupEnv)
want := tc.name + ": " + tc.want
if err == nil || err.Error() != want {
t.Errorf("error %v, want %s", err, want)
}
})
}
}
func TestMaskZoneKeyMasksTheFirstLabelOfAZoneUnderDqSpamhausNet(t *testing.T) {
t.Parallel()
for zone, want := range map[string]string{
spamhaus: spamhausMasked,
spamhaus + ".": spamhausMasked + ".",
"KEY.ZEN.DQ.SPAMHAUS.NET": "********.ZEN.DQ.SPAMHAUS.NET",
dronebl: dronebl,
"dq.spamhaus.net": "dq.spamhaus.net",
spamhaus + ".example": spamhaus + ".example",
} {
if got := config.MaskZoneKey(zone); got != want {
t.Errorf("MaskZoneKey(%q) is %q, want %q", zone, got, want)
}
}
}
func TestDNSBLZoneKeyIsLoggedMaskedAndNeverShown(t *testing.T) {
t.Parallel()
cfg := fromEnvironment(t, environment{dnsblZones: dronebl + ", " + spamhaus})
var out bytes.Buffer
slog.New(slog.NewJSONHandler(&out, nil)).Info("starting", "settings", cfg)
logged := out.String()
if strings.Contains(logged, spamhausKey) ||
!strings.Contains(logged, `"`+dnsblZones+`":"`+dronebl+","+spamhausMasked+`"`) {
t.Errorf("the zones are not logged with the key masked: %s", logged)
}
// Nor does an error that stops the start show a key, in any case.
const (
notZone = " is not a DNS zone such as dnsbl.dronebl.org"
otherKey = "zyxwvutsrqponmlkjihgfedcba"
otherZone = otherKey + ".xbl.dq.spamhaus.net"
)
// 205 characters, 187 with the key masked.
labels := strings.Repeat("a", 63) + "." + strings.Repeat("b", 63) + "." +
strings.Repeat("c", 30) + ".xbl.dq.spamhaus.net"
for _, tc := range []struct{ value, want string }{
{spamhaus + ".", `"` + spamhausMasked + `."` + notZone},
{spamhausKey + "_.xbl.dq.spamhaus.net", `"` + spamhausMasked + `"` + notZone},
{
spamhausKey + "." + labels,
`"********.` + labels + `" is longer than 189 characters, too long ` +
`for the names IPv6 clients are asked about by`,
},
{
spamhaus + "," + strings.ToUpper(spamhaus),
`"********.XBL.DQ.SPAMHAUS.NET" is listed twice`,
},
{spamhaus + "," + otherZone, `"` + spamhausMasked + `" is listed twice`},
{spamhaus + ",,", "has an empty item in its list"},
} {
_, err := config.FromEnvironment(environment{dnsblZones: tc.value}.lookupEnv)
want := dnsblZones + ": " + tc.want
if err == nil || err.Error() != want {
t.Errorf("%s=%s gave the error %v, want %s", dnsblZones, tc.value, err, want)
}
}
}
func TestSizesAndOff(t *testing.T) {
t.Parallel()
@@ -1704,6 +1908,7 @@ func writeFile(t *testing.T, contents string) string {
return path
}
//nolint:funlen // one line for each setting, a list that grows with them
func TestLogsEachSettingWithItsValue(t *testing.T) {
t.Parallel()
@@ -1749,6 +1954,11 @@ func TestLogsEachSettingWithItsValue(t *testing.T) {
blocklistURLs: "",
blocklistRefresh: "24h",
blocklistAction: actionDeny,
dnsblZones: "",
dnsblResolver: "",
reputationAction: "limit:25",
reputationCacheTTL: defaultReputationCacheTTL,
reputationTimeout: "2s",
banResponse: "403",
limitBanDuration: "1h",
limitBanRepeatWindow: "24h",
internal/metrics/metrics.go
+42 -12
View File
@@ -14,6 +14,7 @@ import (
"github.com/prometheus/client_golang/prometheus/promhttp"
"sneak.berlin/go/smallwebwaf/internal/alerts"
"sneak.berlin/go/smallwebwaf/internal/bans"
"sneak.berlin/go/smallwebwaf/internal/config"
"sneak.berlin/go/smallwebwaf/internal/ratelimit"
"sneak.berlin/go/smallwebwaf/internal/remotelog"
"sneak.berlin/go/smallwebwaf/internal/reputation"
@@ -256,24 +257,53 @@ func (m *Metrics) AddLookupFile(lastRead func() time.Time, readFailures func() i
)
}
// AddReputation adds the metrics of the lists fetched from URLs, by
// source, each list's URL: the requests whose client a blocklist lists,
// which ReputationHit counts, and, read from lists as the metrics are
// asked for, the fetches that failed and when the copy in use was fetched.
// It is called once, before ReputationHit.
func (m *Metrics) AddReputation(lists *reputation.Lists) {
// AddReputation adds the metrics of the lists fetched from URLs and of the
// DNSBL zones, by source, each list's URL or each zone, its key masked as
// config.MaskZoneKey masks it: the requests whose client a blocklist or a
// zone's verdict lists, which ReputationHit counts, and, read from lists
// and dnsbl as the metrics are asked for, for a list, the fetches that
// failed and when the copy in use was fetched, and for a zone, the queries
// made and those that failed. It is called once, before ReputationHit.
func (m *Metrics) AddReputation(lists *reputation.Lists, dnsbl *reputation.DNSBL) {
const (
sourceLabel = "source"
failuresHelp = "Fetches of the list, or queries to the DNSBL zone, that failed."
)
m.reputationHits = counterVec("smallwebwaf_reputation_hits_total",
"Requests whose client a blocklist lists, by the blocklist's URL.",
[]string{"source"})
"Requests whose client a blocklist or a DNSBL zone lists, by the "+
"blocklist's URL or the zone.",
[]string{sourceLabel})
m.registry.MustRegister(m.reputationHits)
for _, zone := range dnsbl.Zones() {
source := prometheus.Labels{sourceLabel: config.MaskZoneKey(zone)}
m.registry.MustRegister(
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "smallwebwaf_reputation_queries_total",
Help: "Queries to the DNSBL zone.",
ConstLabels: source,
}, func() float64 {
return float64(dnsbl.Queries(zone))
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "smallwebwaf_reputation_failures_total",
Help: failuresHelp,
ConstLabels: source,
}, func() float64 {
return float64(dnsbl.Failures(zone))
}),
)
}
for _, listURL := range lists.URLs() {
source := prometheus.Labels{"source": listURL}
source := prometheus.Labels{sourceLabel: listURL}
m.registry.MustRegister(
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "smallwebwaf_reputation_failures_total",
Help: "Fetches of the list that failed.",
Help: failuresHelp,
ConstLabels: source,
}, func() float64 {
return float64(lists.Failures(listURL))
@@ -295,8 +325,8 @@ func (m *Metrics) AddReputation(lists *reputation.Lists) {
}
}
// ReputationHit counts a request whose client the blocklist at source, its
// URL, lists.
// ReputationHit counts a request whose client source lists: a blocklist,
// by its URL, or a DNSBL zone, its key masked.
func (m *Metrics) ReputationHit(source string) {
m.reputationHits.WithLabelValues(source).Inc()
}
internal/proxy/biased.go
+22 -15
View File
@@ -28,16 +28,18 @@ func biasedThresholdsSet(cfg *config.Config) bool {
// limitPercentages returns the client's limit percentages, for the rate
// limits and for the byte limits, by its AS number and country as looked
// up, each "" when unknown, and the blocklists that list it. Each is the
// lowest of those the settings give it, the first of them in the order
// below when several are lowest: the percentage SWWAF_ASN_LIMIT_PERCENT
// gives its AS number, the one the file SWWAF_ASN_LIMIT_PERCENT_URL names
// gives it, the one SWWAF_COUNTRY_LIMIT_PERCENT gives its country, for a
// client without a country, SWWAF_UNKNOWN_LIMIT_PERCENT, and for a client
// a blocklist lists, the percentage of SWWAF_BLOCKLIST_ACTION while it is
// limit. For the byte limits, SWWAF_ASN_BYTES_PERCENT and
// SWWAF_COUNTRY_BYTES_PERCENT take the place of the first three for an AS
// number or a country they list.
// up, each "" when unknown, and the blocklists and DNSBL zones that list
// it. Each is the lowest of those the settings give it, the first of them
// in the order below when several are lowest: the percentage
// SWWAF_ASN_LIMIT_PERCENT gives its AS number, the one the file
// SWWAF_ASN_LIMIT_PERCENT_URL names gives it, the one
// SWWAF_COUNTRY_LIMIT_PERCENT gives its country, for a client without a
// country, SWWAF_UNKNOWN_LIMIT_PERCENT, for a client a blocklist lists,
// the percentage of SWWAF_BLOCKLIST_ACTION while it is limit, and for a
// client a DNSBL zone's verdict lists, the percentage of
// SWWAF_REPUTATION_ACTION while it is limit. For the byte limits,
// SWWAF_ASN_BYTES_PERCENT and SWWAF_COUNTRY_BYTES_PERCENT take the place
// of the first three for an AS number or a country they list.
func (rq *request) limitPercentages() (percentage, percentage) {
cfg := rq.h.config
asn, country := rq.line.ASN, rq.line.Country
@@ -52,9 +54,14 @@ func (rq *request) limitPercentages() (percentage, percentage) {
fetched = percentage{percent, "SWWAF_ASN_LIMIT_PERCENT_URL"}
}
listed := percentage{percent: whole}
if len(rq.line.Reputation) > 0 && cfg.BlocklistAction == "limit" {
listed = percentage{cfg.BlocklistLimitPercent, "SWWAF_BLOCKLIST_ACTION"}
blocklisted := percentage{percent: whole}
if rq.blocklisted && cfg.BlocklistAction == "limit" {
blocklisted = percentage{cfg.BlocklistLimitPercent, "SWWAF_BLOCKLIST_ACTION"}
}
dnsblListed := percentage{percent: whole}
if rq.dnsblListed && cfg.ReputationAction == "limit" {
dnsblListed = percentage{cfg.ReputationLimitPercent, "SWWAF_REPUTATION_ACTION"}
}
asnRequests := lowest(given(cfg.ASNLimitPercent, asn, "SWWAF_ASN_LIMIT_PERCENT"),
@@ -71,8 +78,8 @@ func (rq *request) limitPercentages() (percentage, percentage) {
countryBytes = given(cfg.CountryBytesPercent, country, "SWWAF_COUNTRY_BYTES_PERCENT")
}
return lowest(asnRequests, countryRequests, unknown, listed),
lowest(asnBytes, countryBytes, unknown, listed)
return lowest(asnRequests, countryRequests, unknown, blocklisted, dnsblListed),
lowest(asnBytes, countryBytes, unknown, blocklisted, dnsblListed)
}
// given returns the percentage percents, the setting named setting, gives
internal/proxy/proxy.go
+31 -11
View File
@@ -71,15 +71,15 @@ type Params struct {
Rules *rules.Files
// Alerts receive the alert for each ban the proxy makes or makes
// permanent, for each count over an anomaly threshold, for each request
// whose client a blocklist lists, and for GeoJS failing or a fetch of a
// list failing.
// whose client a blocklist or a DNSBL zone lists, and for GeoJS failing,
// a fetch of a list failing or a query to a DNSBL zone failing.
Alerts *alerts.Queue
}
// Server is the server smallwebwaf runs, with the parts of the proxy
// whose state the state files keep, the lookup database, nil unless
// SWWAF_LOOKUP_SOURCE is file, the lists fetched from URLs, which its Run
// fetches, and the metrics.
// fetches, the DNSBL zones' verdicts, and the metrics.
type Server struct {
*http.Server
@@ -89,6 +89,7 @@ type Server struct {
Anomalies *anomaly.Counters
LookupFile *lookup.File
Lists *reputation.Lists
DNSBL *reputation.DNSBL
Metrics *metrics.Metrics
}
@@ -101,6 +102,7 @@ type Server struct {
func New(params Params) *Server {
errorLog := slog.NewLogLogger(params.ProcessLog.Handler(), slog.LevelWarn)
m := metrics.New(params.Config.MetricsTopN, params.Config.InstanceName)
lists, dnsbl := newReputation(params)
h := &handler{
config: params.Config,
requestLog: params.RequestLog,
@@ -136,13 +138,10 @@ func New(params Params) *Server {
Alerts: params.Alerts,
}),
lookupFile: params.LookupFile,
lists: reputation.New(reputation.Params{
BlocklistURLs: params.Config.BlocklistURLs, Refresh: params.Config.BlocklistRefresh,
ASNLimitPercentURL: params.Config.ASNLimitPercentURL, Now: params.Now,
ProcessLog: params.ProcessLog, Alerts: params.Alerts,
}),
rules: params.Rules,
alerts: params.Alerts,
lists: lists,
dnsbl: dnsbl,
rules: params.Rules,
alerts: params.Alerts,
}
h.geojs = lookup.New(lookup.Params{
URL: params.GeoJSURL,
@@ -160,7 +159,7 @@ func New(params Params) *Server {
})
m.AddBansAndClients(h.ledger, h.limiter, params.Now)
m.AddRules(params.Rules)
m.AddReputation(h.lists)
m.AddReputation(h.lists, h.dnsbl)
return &Server{
Server: &http.Server{
@@ -181,10 +180,30 @@ func New(params Params) *Server {
Anomalies: h.anomalies,
LookupFile: h.lookupFile,
Lists: h.lists,
DNSBL: h.dnsbl,
Metrics: m,
}
}
// newReputation returns the lists fetched from URLs and the DNSBL zones'
// verdicts, as the settings in params name them, with none fetched or
// asked for yet.
func newReputation(params Params) (*reputation.Lists, *reputation.DNSBL) {
cfg := params.Config
lists := reputation.New(reputation.Params{
BlocklistURLs: cfg.BlocklistURLs, Refresh: cfg.BlocklistRefresh,
ASNLimitPercentURL: cfg.ASNLimitPercentURL, Now: params.Now,
ProcessLog: params.ProcessLog, Alerts: params.Alerts,
})
dnsbl := reputation.NewDNSBL(reputation.DNSBLParams{
Zones: cfg.DNSBLZones, Resolver: cfg.DNSBLResolver, CacheTTL: cfg.ReputationCacheTTL,
Timeout: cfg.ReputationTimeout, Now: params.Now, ProcessLog: params.ProcessLog,
Alerts: params.Alerts,
})
return lists, dnsbl
}
// handler is the proxy. It holds what every request shares; what belongs
// to one request is in a request.
type handler struct {
@@ -201,6 +220,7 @@ type handler struct {
anomalies *anomaly.Counters
lookupFile *lookup.File
lists *reputation.Lists
dnsbl *reputation.DNSBL
rules *rules.Files
alerts *alerts.Queue
}
internal/proxy/reputation.go
+37 -13
View File
@@ -1,21 +1,47 @@
package proxy
import (
"context"
"sneak.berlin/go/smallwebwaf/internal/alerts"
)
// blocklistDenied notes in the log line the URLs of the blocklists that
// list the client, counts each of them in the metrics and raises a
// reputation_hit alert for it, and reports whether SWWAF_BLOCKLIST_ACTION,
// being deny, refuses the request. Being limit, it lowers the client's
// limits instead (see limitPercentages), and being log, it does nothing
// more.
// blocklistDenied notes the blocklists that list the client, as
// noteListed does, and reports whether SWWAF_BLOCKLIST_ACTION, being deny,
// refuses the request. Being limit, it lowers the client's limits instead
// (see limitPercentages), and being log, it does nothing more.
func (rq *request) blocklistDenied() bool {
listedBy := rq.h.lists.ListedBy(rq.client)
rq.line.Reputation = listedBy
rq.blocklisted = len(listedBy) > 0
rq.noteListed(listedBy, "listed by a blocklist")
for _, listURL := range listedBy {
rq.h.metrics.ReputationHit(listURL)
return rq.blocklisted && rq.h.config.BlocklistAction == "deny"
}
// dnsblDenied notes the DNSBL zones whose verdict lists the client, as
// noteListed does, and reports whether SWWAF_REPUTATION_ACTION, being
// deny, refuses the request. Being limit, it lowers the client's limits
// instead (see limitPercentages), and being log, it does nothing more. A
// zone without a verdict on the client is asked about it in the
// background, and the request does not wait for the answer. ctx is the
// request's own context.
func (rq *request) dnsblDenied(ctx context.Context) bool {
listedBy := rq.h.dnsbl.ListedBy(ctx, rq.client)
rq.dnsblListed = len(listedBy) > 0
rq.noteListed(listedBy, "listed by a DNSBL zone")
return rq.dnsblListed && rq.h.config.ReputationAction == "deny"
}
// noteListed adds sources, the URLs of the blocklists or the DNSBL zones,
// their keys masked, that list the client, to the log line's reputation,
// counts each of them in the metrics, and raises a reputation_hit alert,
// with reason, for each.
func (rq *request) noteListed(sources []string, reason string) {
rq.line.Reputation = append(rq.line.Reputation, sources...)
for _, source := range sources {
rq.h.metrics.ReputationHit(source)
rq.h.alerts.Raise(alerts.Alert{
Event: alerts.EventReputationHit,
Client: rq.client,
@@ -23,10 +49,8 @@ func (rq *request) blocklistDenied() bool {
ASN: rq.line.ASN,
ASName: rq.line.ASName,
Country: rq.line.Country,
Reason: "listed by a blocklist",
Detail: map[string]any{"source": listURL},
Reason: reason,
Detail: map[string]any{"source": source},
})
}
return len(listedBy) > 0 && rq.h.config.BlocklistAction == "deny"
}
internal/proxy/reputation_test.go
+335 -8
View File
@@ -5,6 +5,7 @@ import (
"net/http"
"net/netip"
"slices"
"strings"
"testing"
"time"
@@ -21,11 +22,14 @@ const (
asnLimitPercentURL = "SWWAF_ASN_LIMIT_PERCENT_URL"
)
// The actions of SWWAF_BLOCKLIST_ACTION but limit, which has a
// percentage.
// The actions of SWWAF_BLOCKLIST_ACTION and SWWAF_REPUTATION_ACTION:
// limitHalf gives a listed client half of every limit, and limitQuarter a
// quarter.
const (
actionDeny = "deny"
actionLog = "log"
actionDeny = "deny"
actionLog = "log"
limitHalf = "limit:50"
limitQuarter = "limit:25"
)
// The lists these tests name, which are never fetched: each test puts in
@@ -55,7 +59,7 @@ func TestEachBlocklistActionForAListedAddressAndAListedNetblock(t *testing.T) {
},
{
// Half of 4 requests a minute: the third breaks the limit.
"limit:50", []int{http.StatusOK, http.StatusOK, http.StatusForbidden},
limitHalf, []int{http.StatusOK, http.StatusOK, http.StatusForbidden},
[]string{forward, forward, requestlog.ActionRateLimited},
"50 from " + blocklistAction,
},
@@ -151,10 +155,10 @@ func TestBlocklistLimitTakesPartInTheLowestPercentageOfEveryLimit(t *testing.T)
// percentText gives them, and limitHit its limit_hit.
want, limitHit string
}{
{"limit:50", asnDEQuarter, "25 from " + asnLimitPercent, minuteBytes},
{"limit:25", asnDEHalf, "25 from " + blocklistAction, minuteBytes},
{limitHalf, asnDEQuarter, "25 from " + asnLimitPercent, minuteBytes},
{limitQuarter, asnDEHalf, "25 from " + blocklistAction, minuteBytes},
// The AS number's, the first of two alike.
{"limit:25", asnDEQuarter, "25 from " + asnLimitPercent, minuteBytes},
{limitQuarter, asnDEQuarter, "25 from " + asnLimitPercent, minuteBytes},
{actionLog, asnDE + ":100", none, ""},
} {
t.Run(tc.action+" "+tc.asnPercent, func(t *testing.T) {
@@ -299,11 +303,334 @@ func TestEachBlocklistThatListsAClientRaisesAnAlertOncePerCooldownAndIsCounted(
}
}
// The DNSBL settings.
const (
dnsblZones = "SWWAF_DNSBL_ZONES"
dnsblResolver = "SWWAF_DNSBL_RESOLVER"
reputationAction = "SWWAF_REPUTATION_ACTION"
)
// The DNSBL zones these tests name, which are never asked about the
// clients the tests send requests from: each test puts in the verdicts it
// needs, as reputation.json would at start. A query a test does start is
// sent to noResolver, where nothing listens, so that none leaves the host.
const (
dnsblZone = "dnsbl.example"
otherZone = "other.example"
noResolver = "127.0.0.1:9"
)
func TestEachReputationActionForAClientADNSBLZoneLists(t *testing.T) {
t.Parallel()
forward, denied := requestlog.ActionForward, requestlog.ActionDenied
for _, tc := range []struct {
action string
// statuses and actions are those of a listed client's three
// requests, and percent their limit_percent, as percentText gives it.
statuses []int
actions []string
percent string
}{
{
actionDeny, []int{http.StatusForbidden, http.StatusForbidden, http.StatusForbidden},
[]string{denied, denied, denied}, none,
},
{
// Half of 4 requests a minute: the third breaks the limit.
limitHalf, []int{http.StatusOK, http.StatusOK, http.StatusForbidden},
[]string{forward, forward, requestlog.ActionRateLimited},
"50 from " + reputationAction,
},
{
actionLog, []int{http.StatusOK, http.StatusOK, http.StatusOK},
[]string{forward, forward, forward}, none,
},
} {
t.Run(tc.action, func(t *testing.T) {
t.Parallel()
s, server, _ := startWithLookups(t, map[string]string{
rateLimitPerMinute: fourAMinute, dnsblZones: dnsblZone + "," + otherZone,
dnsblResolver: noResolver, reputationAction: tc.action,
})
listedBy := map[string][]string{
fromDE: {dnsblZone, otherZone}, fromKP: {otherZone}, unplaced: nil,
}
loadVerdicts(server, listedBy)
for _, from := range []string{fromDE, fromKP} {
for i := range 3 {
line := s.get(from, tc.statuses[i], tc.actions[i])
// In the order SWWAF_DNSBL_ZONES names them.
wantReputation(t, line, listedBy[from]...)
wantPercent(t, "limit_percent", line.LimitPercent,
line.LimitPercentSetting, tc.percent)
// A request refused for the verdict is not counted.
counted := line.fields["counts"] != nil
if counted != (tc.actions[i] != denied) {
t.Errorf("request from %s counted %t, logged %s", from, counted,
tc.actions[i])
}
}
}
// A client no zone lists has the whole limit.
for range 3 {
line := s.get(unplaced, http.StatusOK, forward)
wantReputation(t, line)
wantPercent(t, "limit_percent", line.LimitPercent, line.LimitPercentSetting,
none)
}
// A refusal for the verdict makes no ban, and every client had its
// verdicts, so no zone was asked.
if held := server.Ledger.Snapshot(); tc.action == actionDeny && len(held) != 0 {
t.Errorf("bans %+v, want none", held)
}
if queries := server.DNSBL.Queries(dnsblZone); queries != 0 {
t.Errorf("%d queries, want none", queries)
}
})
}
}
func TestDNSBLZonesComeAfterTheBlocklistsAndSkipAllowNets(t *testing.T) {
t.Parallel()
s, server, queue := startWithLookups(t, map[string]string{
blocklistURLs: dropURL, dnsblZones: dnsblZone, dnsblResolver: noResolver,
reputationAction: actionDeny, allowNets: fromDE,
})
loadLists(t, server, map[string][]string{dropURL: {fromKP}})
loadVerdicts(server, map[string][]string{fromKP: {dnsblZone}, fromDE: {dnsblZone}})
// The blocklist refuses fromKP before its verdict is looked at, and
// fromDE, in SWWAF_ALLOW_NETS, is not checked at all: neither is noted
// for the zone, nor alerted, nor asked about.
wantReputation(t, s.get(fromKP, http.StatusForbidden, requestlog.ActionDenied),
dropURL)
wantReputation(t, s.get(fromDE, http.StatusOK, requestlog.ActionForward))
waiting := queue.Snapshot().Waiting[alerts.DestinationWebhook]
if len(waiting) != 1 || waiting[0].Detail["source"] != dropURL {
t.Errorf("alerts waiting %+v, want the blocklist's reputation_hit alone", waiting)
}
if queries := server.DNSBL.Queries(dnsblZone); queries != 0 {
t.Errorf("%d queries, want none", queries)
}
}
func TestObserveModeForwardsAClientADNSBLZoneDeniesAndAlertsIt(t *testing.T) {
t.Parallel()
s, server, queue := startWithLookups(t, map[string]string{
dnsblZones: dnsblZone, dnsblResolver: noResolver, reputationAction: actionDeny,
mode: observe,
})
loadVerdicts(server, map[string][]string{fromDE: {dnsblZone}})
line := s.get(fromDE, http.StatusOK, requestlog.ActionForward)
wantWouldAction(t, line, requestlog.ActionDenied)
wantReputation(t, line, dnsblZone)
waiting := queue.Snapshot().Waiting[alerts.DestinationWebhook]
if len(waiting) != 1 || waiting[0].Event != alerts.EventReputationHit {
t.Errorf("alerts waiting %+v, want a reputation_hit alert", waiting)
}
}
func TestReputationLimitTakesPartInTheLowestPercentageOfEveryLimit(t *testing.T) {
t.Parallel()
for _, tc := range []struct {
blocklistAction, reputationAction string
// want is the request's limit_percent and bytes_percent, as
// percentText gives them.
want string
}{
{limitHalf, limitQuarter, "25 from " + reputationAction},
{limitQuarter, limitHalf, "25 from " + blocklistAction},
// The blocklist's, the first of two alike.
{limitQuarter, limitQuarter, "25 from " + blocklistAction},
{actionLog, limitQuarter, "25 from " + reputationAction},
{actionLog, actionLog, none},
} {
t.Run(tc.blocklistAction+" "+tc.reputationAction, func(t *testing.T) {
t.Parallel()
s, server, _ := startWithLookups(t, map[string]string{
blocklistURLs: dropURL, blocklistAction: tc.blocklistAction,
dnsblZones: dnsblZone, dnsblResolver: noResolver,
reputationAction: tc.reputationAction,
})
loadLists(t, server, map[string][]string{dropURL: {fromDE}})
loadVerdicts(server, map[string][]string{fromDE: {dnsblZone}})
// Named by the blocklist, then by the zone.
line := s.get(fromDE, http.StatusOK, requestlog.ActionForward)
wantReputation(t, line, dropURL, dnsblZone)
wantPercent(t, "limit_percent", line.LimitPercent, line.LimitPercentSetting,
tc.want)
wantPercent(t, "bytes_percent", line.BytesPercent, line.BytesPercentSetting,
tc.want)
})
}
}
func TestEachZoneThatListsAClientRaisesAnAlertOncePerCooldownAndIsCounted(
t *testing.T,
) {
t.Parallel()
s, clk, server, queue := startWithLookupsAndClock(t, map[string]string{
dnsblZones: dnsblZone + "," + otherZone, dnsblResolver: noResolver,
reputationAction: actionLog, metricsToken: token,
})
loadVerdicts(server, map[string][]string{
fromDE: {dnsblZone, otherZone}, unplaced: nil,
})
// The second request's alerts are repeats, which the cooldown holds
// back.
for range 2 {
wantReputation(t, s.get(fromDE, http.StatusOK, requestlog.ActionForward),
dnsblZone, otherZone)
}
hit := func(zone string) alerts.Alert {
return alerts.Alert{
Instance: alertInstance,
Time: clk.Now(),
Event: alerts.EventReputationHit,
Client: netip.MustParseAddr(fromDE),
Netblock: netip.MustParsePrefix(fromDE + "/32"),
ASN: asnDE,
ASName: asNameDE,
Country: "DE",
Reason: "listed by a DNSBL zone",
Detail: map[string]any{"source": zone},
}
}
wantAlerts(t, queue, hit(dnsblZone), hit(otherZone))
if queue.Suppressed() != 2 {
t.Errorf("%d alerts held back, want the second request's 2", queue.Suppressed())
}
// Each zone's hits, and its queries and their failures, none, since
// every client had its verdicts.
metrics := s.scrape(unplaced)
for _, zone := range []string{dnsblZone, otherZone} {
labels := `{instance="` + alertInstance + `",source="` + zone + `"}`
wantMetric(t, metrics, "smallwebwaf_reputation_hits_total"+labels, 2)
wantMetric(t, metrics, "smallwebwaf_reputation_queries_total"+labels, 0)
wantMetric(t, metrics, "smallwebwaf_reputation_failures_total"+labels, 0)
}
}
func TestZoneKeyIsMaskedInTheLogTheAlertAndTheMetrics(t *testing.T) {
t.Parallel()
const (
key = "abcdefghijklmnopqrstuvwxyz"
keyed = key + ".xbl.dq.spamhaus.net"
masked = "********.xbl.dq.spamhaus.net"
)
s, server, queue := startWithLookups(t, map[string]string{
dnsblZones: keyed, dnsblResolver: noResolver, reputationAction: actionLog,
metricsToken: token,
})
loadVerdicts(server, map[string][]string{fromDE: {keyed}, unplaced: nil})
wantReputation(t, s.get(fromDE, http.StatusOK, requestlog.ActionForward), masked)
waiting := queue.Snapshot().Waiting[alerts.DestinationWebhook]
if len(waiting) != 1 || waiting[0].Detail["source"] != masked {
t.Errorf("alerts waiting %+v, want a reputation_hit alert from %s", waiting,
masked)
}
metrics := s.scrape(unplaced)
wantMetric(t, metrics, `smallwebwaf_reputation_hits_total{instance="`+
alertInstance+`",source="`+masked+`"}`, 1)
for name, shown := range map[string]string{
"the log": s.out.text(), "the metrics": metrics,
} {
if strings.Contains(shown, key) {
t.Errorf("%s shows the key:\n%s", name, shown)
}
}
}
func TestRequestFromAClientWithoutAVerdictHasTheZoneAskedAboutIt(t *testing.T) {
t.Parallel()
s, server, _ := startWithLookups(t, map[string]string{
dnsblZones: dnsblZone + "," + otherZone, dnsblResolver: noResolver,
})
server.DNSBL.Load([]reputation.Verdict{{
Zone: otherZone, Client: netip.MustParseAddr(fromDE), Listed: true,
Fetched: verdictsFetched(),
}})
// The verdict of the other zone is used, and dnsbl.example, which has
// none, is asked about the client in the background, once: the second
// request finds the query under way, or the zone left alone after it
// failed, since nothing answers at noResolver.
for range 2 {
wantReputation(t, s.get(fromDE, http.StatusOK, requestlog.ActionForward), otherZone)
}
if queries := server.DNSBL.Queries(dnsblZone); queries != 1 {
t.Errorf("%d queries to %s, want 1", queries, dnsblZone)
}
if queries := server.DNSBL.Queries(otherZone); queries != 0 {
t.Errorf("%d queries to %s, want none", queries, otherZone)
}
}
// listsFetched is when loadLists has the copies fetched.
func listsFetched() time.Time {
return time.Date(2026, 10, 5, 0, 0, 0, 0, time.UTC)
}
// verdictsFetched is when loadVerdicts has the verdicts fetched: half a
// day before the time the tests' clock is set to, so that they are in use
// until half a day later.
func verdictsFetched() time.Time {
return time.Date(2026, 10, 5, 12, 0, 0, 0, time.UTC)
}
// loadVerdicts puts into server's DNSBL, for each client listedBy names,
// a verdict of each zone SWWAF_DNSBL_ZONES names, fetched at
// verdictsFetched, as reputation.json would at start: one that lists the
// client from each zone listedBy gives for it, and one that does not from
// each other zone.
func loadVerdicts(server *proxy.Server, listedBy map[string][]string) {
verdicts := make([]reputation.Verdict, 0, len(listedBy)*len(server.DNSBL.Zones()))
for client, zones := range listedBy {
for _, zone := range server.DNSBL.Zones() {
verdicts = append(verdicts, reputation.Verdict{
Zone: zone, Client: netip.MustParseAddr(client),
Listed: slices.Contains(zones, zone), Fetched: verdictsFetched(),
})
}
}
server.DNSBL.Load(verdicts)
}
// loadLists puts copies of lists into server's lists, by URL, each with
// its lines, fetched at listsFetched, as reputation.json would at start.
func loadLists(t *testing.T, server *proxy.Server, copies map[string][]string) {
internal/proxy/request.go
+16 -9
View File
@@ -63,7 +63,10 @@ type request struct {
// limits and for the byte limits.
counted bool
limitPercent, bytesPercent percentage
start time.Time
// blocklisted is true once a blocklist is found to list the client,
// and dnsblListed once a DNSBL zone's verdict is.
blocklisted, dnsblListed bool
start time.Time
// checked is when the checks were done, and upstreamStart when the
// request was handed to the app.
checked time.Time
@@ -213,14 +216,14 @@ func (rq *request) check(ctx context.Context) *refusal {
// client in SWWAF_ALLOW_NETS skips them, and is not looked up. For any
// other client, SWWAF_DENY_NETS comes first, then a ban on its netblock,
// so that a client either refuses is not looked up, then the lookup of
// its AS number and country, then the country lists, and then the
// blocklists; a request any of them refuses is not counted for the rate
// limits. Then come the rate limits, unless the client is in
// SWWAF_RATE_LIMIT_EXEMPT_NETS or the request's path is exempt under
// SWWAF_RATE_LIMIT_EXEMPT_PATHS, so that every other request is counted,
// each of them by the client's limit percentages, and last the rule
// files. A request exempt from the rate limits is exempt from the byte
// limits too. ctx is the request's own context.
// its AS number and country, then the country lists, then the blocklists,
// and then the DNSBL zones' verdicts; a request any of them refuses is not
// counted for the rate limits. Then come the rate limits, unless the
// client is in SWWAF_RATE_LIMIT_EXEMPT_NETS or the request's path is
// exempt under SWWAF_RATE_LIMIT_EXEMPT_PATHS, so that every other request
// is counted, each of them by the client's limit percentages, and last the
// rule files. A request exempt from the rate limits is exempt from the
// byte limits too. ctx is the request's own context.
func (rq *request) checkClient(ctx context.Context) string {
cfg := rq.h.config
if isInside(rq.client, cfg.AllowNets) {
@@ -247,6 +250,10 @@ func (rq *request) checkClient(ctx context.Context) string {
return requestlog.ActionDenied
}
if rq.dnsblDenied(ctx) {
return requestlog.ActionDenied
}
rq.counted = !isInside(rq.client, cfg.RateLimitExemptNets) &&
!pathExempt(rq.in.URL, cfg.RateLimitExemptPaths)
if rq.counted {
internal/reputation/dnsbl.go
+335
View File
@@ -0,0 +1,335 @@
package reputation
import (
"cmp"
"context"
"encoding/hex"
"errors"
"fmt"
"log/slog"
"net"
"net/netip"
"slices"
"strings"
"sync"
"time"
"github.com/hashicorp/golang-lru/v2/simplelru"
"sneak.berlin/go/smallwebwaf/internal/alerts"
"sneak.berlin/go/smallwebwaf/internal/config"
)
const (
// maxVerdicts is how many verdicts are kept. Past it, the one fetched
// longest ago is dropped.
maxVerdicts = 100000
// maxQueries is how many queries may be under way at once. Past it, a
// zone is not asked about a client until the client's next request, so
// that a swarm of new addresses cannot fill the memory.
maxQueries = 1000
// failureDelay is how long a zone is not asked again after a query to
// it fails, so that a zone refusing queries is not asked on every
// request.
failureDelay = time.Minute
)
var (
errAsk = errors.New("ask the zone")
errRefused = errors.New("the zone refused the query")
errNotListing = errors.New("the answer is outside 127.0.0.0/8")
)
// Verdict is what a zone said about a client, as reputation.json holds
// it: the zone, the client's address, whether the zone lists it, and when
// the zone answered.
type Verdict struct {
Zone string `json:"zone"`
Client netip.Addr `json:"client"`
Listed bool `json:"listed"`
Fetched time.Time `json:"fetched"`
}
// DNSBLParams are what NewDNSBL needs.
type DNSBLParams struct {
// Zones are the DNSBL zones clients are asked about in
// (SWWAF_DNSBL_ZONES).
Zones []string
// Resolver is the resolver they are asked through
// (SWWAF_DNSBL_RESOLVER), or, while it is the zero AddrPort, the
// host's, as /etc/resolv.conf names it.
Resolver netip.AddrPort
// CacheTTL is how long a verdict is used after it was fetched
// (SWWAF_REPUTATION_CACHE_TTL), and Timeout how long a query may take
// (SWWAF_REPUTATION_TIMEOUT).
CacheTTL time.Duration
Timeout time.Duration
// Now tells the time, normally time.Now in UTC.
Now func() time.Time
// ProcessLog receives each query that fails, and why.
ProcessLog *slog.Logger
// Alerts receive a source_failure alert for each query that fails.
Alerts *alerts.Queue
}
// DNSBL asks the DNSBL zones about clients, in the background, and keeps
// their verdicts. It is safe for concurrent use.
type DNSBL struct {
params DNSBLParams
resolver *net.Resolver
mu sync.Mutex
// verdicts are by query. Each is added as it is fetched and never moved
// up, so that the one fetched longest ago is the first dropped.
verdicts *simplelru.LRU[query, Verdict]
// asking are the queries under way.
asking map[query]bool
// queries and failures count, by zone, the queries made and those that
// failed, and retryAt is when a zone whose last query failed may be
// asked again.
queries map[string]int
failures map[string]int
retryAt map[string]time.Time
}
// query is a client's address, to ask a zone about.
type query struct {
zone string
client netip.Addr
}
// NewDNSBL returns a DNSBL with no verdict yet.
func NewDNSBL(params DNSBLParams) *DNSBL {
verdicts, err := simplelru.NewLRU[query, Verdict](maxVerdicts, nil)
if err != nil {
panic(err) // NewLRU fails only for a size below one
}
resolver := &net.Resolver{}
if params.Resolver.IsValid() {
// Dial is used by Go's own resolver alone.
resolver.PreferGo = true
resolver.Dial = func(ctx context.Context, network, _ string) (net.Conn, error) {
var dialer net.Dialer
return dialer.DialContext(ctx, network, params.Resolver.String())
}
}
return &DNSBL{
params: params,
resolver: resolver,
verdicts: verdicts,
asking: map[query]bool{},
queries: map[string]int{},
failures: map[string]int{},
retryAt: map[string]time.Time{},
}
}
// Zones returns the zones, in the order SWWAF_DNSBL_ZONES names them.
func (d *DNSBL) Zones() []string {
return slices.Clone(d.params.Zones)
}
// ListedBy returns the zones whose verdict on addr, a client's address,
// lists it, in the order SWWAF_DNSBL_ZONES names them, each with its key
// masked, as config.MaskZoneKey masks it, since they go to the request
// log, the alerts and the metrics. A verdict is used until CacheTTL has
// passed since it was fetched. Each zone without one is asked about addr
// in the background, unless a query about addr to it is under way, the
// zone is left alone after a failure, or maxQueries are under way;
// ListedBy never waits for a query. ctx is the context of the client's
// request, and a query goes on after the request ends.
func (d *DNSBL) ListedBy(ctx context.Context, addr netip.Addr) []string {
d.mu.Lock()
defer d.mu.Unlock()
now := d.params.Now()
var listedBy []string
for _, zone := range d.params.Zones {
q := query{zone: zone, client: addr}
kept, found := d.verdicts.Peek(q)
switch {
case found && now.Sub(kept.Fetched) < d.params.CacheTTL:
if kept.Listed {
listedBy = append(listedBy, config.MaskZoneKey(zone))
}
case !d.asking[q] && !now.Before(d.retryAt[zone]) && len(d.asking) < maxQueries:
d.asking[q] = true
d.queries[zone]++
go d.ask(context.WithoutCancel(ctx), q)
}
}
return listedBy
}
// Queries returns how many queries were made to zone.
func (d *DNSBL) Queries(zone string) int {
d.mu.Lock()
defer d.mu.Unlock()
return d.queries[zone]
}
// Failures returns how many queries to zone failed.
func (d *DNSBL) Failures(zone string) int {
d.mu.Lock()
defer d.mu.Unlock()
return d.failures[zone]
}
// Snapshot returns every verdict still in use, sorted by client, then by
// zone, as reputation.json lists them.
func (d *DNSBL) Snapshot() []Verdict {
d.mu.Lock()
now := d.params.Now()
verdicts := make([]Verdict, 0, d.verdicts.Len())
for _, kept := range d.verdicts.Values() {
if now.Sub(kept.Fetched) < d.params.CacheTTL {
verdicts = append(verdicts, kept)
}
}
d.mu.Unlock()
slices.SortFunc(verdicts, func(a, b Verdict) int {
return cmp.Or(a.Client.Compare(b.Client), strings.Compare(a.Zone, b.Zone))
})
return verdicts
}
// Load keeps verdicts, read from reputation.json, in place of those it
// keeps, but for those of a zone SWWAF_DNSBL_ZONES does not name, and,
// past maxVerdicts, those fetched longest ago. One fetched CacheTTL ago or
// more is neither used nor written, as for any verdict.
func (d *DNSBL) Load(verdicts []Verdict) {
verdicts = slices.Clone(verdicts)
slices.SortStableFunc(verdicts, func(a, b Verdict) int {
return a.Fetched.Compare(b.Fetched)
})
d.mu.Lock()
defer d.mu.Unlock()
d.verdicts.Purge()
for _, kept := range verdicts {
if slices.Contains(d.params.Zones, kept.Zone) {
d.verdicts.Add(query{zone: kept.Zone, client: kept.Client}, kept)
}
}
}
// ask asks q's zone about q's client, keeps the verdict, and notes the
// query as no longer under way. A query that fails gives no verdict: it
// is counted, logged and raised as a source_failure alert, which show the
// zone with its key masked, and the zone is not asked again for
// failureDelay.
func (d *DNSBL) ask(ctx context.Context, q query) {
listed, err := d.lookUp(ctx, q)
now := d.params.Now()
d.mu.Lock()
delete(d.asking, q)
if err == nil {
d.verdicts.Add(q, Verdict{
Zone: q.zone, Client: q.client, Listed: listed, Fetched: now,
})
} else {
d.failures[q.zone]++
d.retryAt[q.zone] = now.Add(failureDelay)
}
d.mu.Unlock()
if err != nil {
const failed = "asking a DNSBL zone failed"
shown := config.MaskZoneKey(q.zone)
// Raised before it is logged, so that the alert is there once the
// log line is.
d.params.Alerts.Raise(alerts.Alert{
Event: alerts.EventSourceFailure,
Reason: failed,
Detail: map[string]any{"source": shown, "error": err.Error()},
})
d.params.ProcessLog.Warn(failed, "zone", shown, "error", err.Error())
}
}
// lookUp asks q's zone about q's client through the resolver, and returns
// whether the zone lists it, as readAnswer reads the answer. No such name
// is a client the zone does not list. A query not answered within Timeout
// fails.
func (d *DNSBL) lookUp(ctx context.Context, q query) (bool, error) {
ctx, cancel := context.WithTimeout(ctx, d.params.Timeout)
defer cancel()
answer, err := d.resolver.LookupNetIP(ctx, "ip4", queryName(q.zone, q.client))
var dnsErr *net.DNSError
switch {
case err == nil:
return readAnswer(answer)
case errors.As(err, &dnsErr) && dnsErr.IsNotFound:
return false, nil
case errors.As(err, &dnsErr):
// The error names the name asked about, which holds the client's
// address, which is not to be logged: only what went wrong is kept.
return false, fmt.Errorf("%w: %s", errAsk, dnsErr.Err)
default:
return false, fmt.Errorf("%w: %w", errAsk, err)
}
}
// queryName returns the name a zone is asked about addr by, as RFC 5782
// builds it: the four numbers of an IPv4 address, or the 32 hex digits of
// an IPv6 address, in reverse order, each followed by a dot, then the zone
// and a dot, which makes it a full name, to which the resolver adds no
// search domain of /etc/resolv.conf.
func queryName(zone string, addr netip.Addr) string {
parts := strings.Split(addr.String(), ".")
if addr.Is6() {
parts = strings.Split(hex.EncodeToString(addr.AsSlice()), "")
}
slices.Reverse(parts)
return strings.Join(parts, ".") + "." + zone + "."
}
// readAnswer reads the addresses a zone answered with. An address in
// 127.0.0.0/8 lists the client, as RFC 5782 has zones answer, but one in
// 127.255.255.0/24 is how Spamhaus refuses a query, such as one sent
// through a public resolver or one past its limit, and is a failure. So is
// an address outside 127.0.0.0/8, such as a resolver gives that answers
// even for names that do not exist.
func readAnswer(answer []netip.Addr) (bool, error) {
listing := netip.MustParsePrefix("127.0.0.0/8")
refusal := netip.MustParsePrefix("127.255.255.0/24")
for _, addr := range answer {
switch {
case refusal.Contains(addr):
return false, fmt.Errorf("%w: %s", errRefused, addr)
case !listing.Contains(addr):
return false, fmt.Errorf("%w: %s", errNotListing, addr)
}
}
return len(answer) > 0, nil
}
internal/reputation/dnsbl_test.go
+737
View File
@@ -0,0 +1,737 @@
package reputation_test
import (
"bytes"
"context"
"encoding/binary"
"errors"
"io"
"log/slog"
"net"
"net/http"
"net/http/httptest"
"net/netip"
"reflect"
"slices"
"strings"
"sync"
"testing"
"testing/synctest"
"time"
"sneak.berlin/go/smallwebwaf/internal/alerts"
"sneak.berlin/go/smallwebwaf/internal/metrics"
"sneak.berlin/go/smallwebwaf/internal/reputation"
)
// The tests of the DNSBL zones run in synctest bubbles, as those of the
// lists do, and the resolver the zones are asked through is a stand-in
// reached through an in-memory connection, net.Pipe's, for the same
// reason. They run one at a time, none in parallel with another test of
// this package: Go's resolver counts the queries under way in one
// sync.WaitGroup for the whole process, and the process fails when
// queries from two bubbles, or from a bubble and from outside one, are
// under way at once. TestMain has the resolver make its configuration,
// which it makes on its first query, outside every bubble, since the
// configuration holds a channel, which the bubble it was made in would
// keep to itself.
const (
// zone and otherZone are the DNSBL zones the tests name.
zone = "dnsbl.example"
otherZone = "other.example"
// cacheTTL is the tests' SWWAF_REPUTATION_CACHE_TTL, and timeout their
// SWWAF_REPUTATION_TIMEOUT: a second, the least time /etc/resolv.conf
// can have Go's resolver wait for one server, so that it is the
// DNSBL's own timeout that ends a query, whatever that file says.
cacheTTL = 24 * time.Hour
timeout = time.Second
// listed and unlisted are clients zone is asked about by the names
// listedName and unlistedName, and most tests have zone list the first
// alone, by answering with listing.
listed = "192.0.2.99"
unlisted = "192.0.2.100"
listedName = "99.2.0.192." + zone + "."
unlistedName = "100.2.0.192." + zone + "."
listing = "127.0.0.2"
)
// The DNS response codes the stand-in answers with, besides no error.
const (
serverFailure = 2
noSuchName = 3
refused = 5
)
var errNoNetwork = errors.New("the test dials nothing")
func TestMain(m *testing.M) {
// A query that fails at once, as nothing is dialled for it.
resolver := &net.Resolver{
PreferGo: true,
Dial: func(context.Context, string, string) (net.Conn, error) {
return nil, errNoNetwork
},
}
_, _ = resolver.LookupNetIP(context.Background(), "ip4", "warm-up.invalid.")
m.Run()
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestZonesListOrNotClientsByTheirIPv4AndIPv6Addresses(t *testing.T) {
synctest.Test(t, func(t *testing.T) {
// The addresses of the examples of RFC 5782, and the names it
// gives for them.
const (
v4 = "192.0.2.99"
v6 = "2001:db8:1:2:3:4:567:89ab"
// v6Name is the hex digits of v6, in reverse order.
v6Name = "b.a.9.8.7.6.5.0.4.0.0.0.3.0.0.0.2.0.0.0.1.0.0.0.8.b.d.0.1.0.0.2."
)
resolver := &resolverStandIn{answers: map[string]answer{
"99.2.0.192." + zone + ".": {addrs: []string{listing}},
v6Name + otherZone + ".": {addrs: []string{"127.0.0.4", "127.0.0.10"}},
"99.2.0.192." + otherZone + ".": {},
}}
dnsbl := newDNSBL(resolver, dnsblParams(zone, otherZone))
// Neither client has a verdict yet, so neither is listed, and each
// zone is asked about each.
wantZones(t, dnsbl, v4)
wantZones(t, dnsbl, v6)
synctest.Wait()
wantZones(t, dnsbl, v4, zone)
wantZones(t, dnsbl, v6, otherZone)
wantAsked(t, resolver,
"99.2.0.192."+zone+".", "99.2.0.192."+otherZone+".",
v6Name+zone+".", v6Name+otherZone+".")
})
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestListedByNeverWaitsForAQuery(t *testing.T) {
synctest.Test(t, func(t *testing.T) {
dnsbl := newDNSBL(&resolverStandIn{hanging: true}, dnsblParams(zone))
began := time.Now()
// The second, while the first's query is under way, starts none.
wantZones(t, dnsbl, listed)
wantZones(t, dnsbl, listed)
if waited := time.Since(began); waited != 0 {
t.Errorf("waited %s for the query, want no wait", waited)
}
synctest.Wait()
wantQueries(t, dnsbl, 1, 0)
waitForTheResolver()
})
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestVerdictUsedUntilTheCacheTTLHasPassedSinceItWasFetched(t *testing.T) {
synctest.Test(t, func(t *testing.T) {
resolver := &resolverStandIn{answers: map[string]answer{
listedName: {addrs: []string{listing}},
}}
dnsbl := newDNSBL(resolver, dnsblParams(zone))
wantZones(t, dnsbl, listed)
wantZones(t, dnsbl, unlisted)
synctest.Wait()
// The zone lists the other client from now on, but the verdicts
// kept are used, and the zone is not asked again, until the TTL
// has passed.
resolver.set(listedName, answer{rcode: noSuchName})
resolver.set(unlistedName, answer{addrs: []string{listing}})
time.Sleep(cacheTTL - time.Nanosecond)
wantZones(t, dnsbl, listed, zone)
wantZones(t, dnsbl, unlisted)
synctest.Wait()
wantQueries(t, dnsbl, 2, 0)
// Then neither verdict is used, and both clients are asked about
// again.
time.Sleep(time.Nanosecond)
wantZones(t, dnsbl, listed)
wantZones(t, dnsbl, unlisted)
synctest.Wait()
wantQueries(t, dnsbl, 4, 0)
wantZones(t, dnsbl, listed)
wantZones(t, dnsbl, unlisted, zone)
})
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestQueryNotAnsweredWithinTheTimeoutFails(t *testing.T) {
synctest.Test(t, func(t *testing.T) {
queue := newQueue()
p := dnsblParams(zone)
p.Alerts = queue
dnsbl := newDNSBL(&resolverStandIn{hanging: true}, p)
wantZones(t, dnsbl, listed)
time.Sleep(timeout - time.Nanosecond)
synctest.Wait()
wantQueries(t, dnsbl, 1, 0)
time.Sleep(time.Nanosecond)
synctest.Wait()
wantQueries(t, dnsbl, 1, 1)
if got := waiting(queue); len(got) != 1 ||
got[0].Detail["error"] != "ask the zone: i/o timeout" {
t.Errorf("alerts waiting %+v, want the timeout's", got)
}
if verdicts := dnsbl.Snapshot(); len(verdicts) != 0 {
t.Errorf("verdicts %+v, want none", verdicts)
}
waitForTheResolver()
})
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestZoneThatFailsOrRefusesGivesNoVerdictAndIsLeftAloneForAMinute(t *testing.T) {
for _, tc := range []struct {
name string
answer answer
error string
}{
{
"a server failure", answer{rcode: serverFailure},
"ask the zone: server misbehaving",
},
{"a refusal", answer{rcode: refused}, "ask the zone: server misbehaving"},
{
"an answer in 127.255.255.0/24, with which Spamhaus refuses a query",
answer{addrs: []string{"127.255.255.254"}},
"the zone refused the query: 127.255.255.254",
},
{
"an answer outside 127.0.0.0/8, as for a name that does not exist",
answer{addrs: []string{"192.0.2.1"}},
"the answer is outside 127.0.0.0/8: 192.0.2.1",
},
} {
//nolint:paralleltest // one at a time, as the comment at the top of this file says
t.Run(tc.name, func(t *testing.T) {
synctest.Test(t, func(t *testing.T) {
var log bytes.Buffer
queue := newQueue()
p := dnsblParams(zone)
p.Alerts = queue
p.ProcessLog = slog.New(slog.NewJSONHandler(&log, nil))
dnsbl := newDNSBL(&resolverStandIn{answers: map[string]answer{
listedName: tc.answer,
}}, p)
// The failure gives no verdict, and the zone is not asked
// again within a minute of it.
wantZones(t, dnsbl, listed)
synctest.Wait()
time.Sleep(time.Minute - time.Nanosecond)
wantZones(t, dnsbl, listed)
synctest.Wait()
wantQueries(t, dnsbl, 1, 1)
time.Sleep(time.Nanosecond)
wantZones(t, dnsbl, listed)
synctest.Wait()
wantQueries(t, dnsbl, 2, 2)
if verdicts := dnsbl.Snapshot(); len(verdicts) != 0 {
t.Errorf("verdicts %+v, want none", verdicts)
}
// One alert for the first failure; the cooldown holds back
// the second.
wantAlert(t, queue, alerts.Alert{
Time: time.Now().Add(-time.Minute),
Event: alerts.EventSourceFailure,
Reason: "asking a DNSBL zone failed",
Detail: map[string]any{"source": zone, "error": tc.error},
})
if !strings.Contains(log.String(), `"msg":"asking a DNSBL zone failed",`+
`"zone":"`+zone+`","error":"`+tc.error) {
t.Errorf("logged\n%s\nwant the failures", log.String())
}
})
})
}
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestAtMost1000QueriesUnderWay(t *testing.T) {
synctest.Test(t, func(t *testing.T) {
dnsbl := newDNSBL(&resolverStandIn{hanging: true}, dnsblParams(zone))
client := netip.MustParseAddr("198.18.0.0")
for range 1001 {
dnsbl.ListedBy(t.Context(), client)
client = client.Next()
}
synctest.Wait()
wantQueries(t, dnsbl, 1000, 0)
waitForTheResolver()
})
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestMetricsCountEachZonesQueriesAndThoseThatFailed(t *testing.T) {
synctest.Test(t, func(t *testing.T) {
resolver := &resolverStandIn{answers: map[string]answer{
listedName: {addrs: []string{listing}},
"99.2.0.192." + otherZone + ".": {rcode: serverFailure},
}}
dnsbl := newDNSBL(resolver, dnsblParams(zone, otherZone))
m := metrics.New(1, "app")
m.AddReputation(reputation.New(params()), dnsbl)
wantZones(t, dnsbl, listed)
synctest.Wait()
scraped := httptest.NewRecorder()
m.ServeHTTP(scraped, httptest.NewRequestWithContext(t.Context(), http.MethodGet,
"/", http.NoBody))
for series, want := range map[string]string{
"queries_total" + `{instance="app",source="` + zone + `"}`: "1",
"failures_total" + `{instance="app",source="` + zone + `"}`: "0",
"queries_total" + `{instance="app",source="` + otherZone + `"}`: "1",
"failures_total" + `{instance="app",source="` + otherZone + `"}`: "1",
} {
line := "\nsmallwebwaf_reputation_" + series + " " + want + "\n"
if !strings.Contains(scraped.Body.String(), line) {
t.Errorf("metrics\n%s\nwant%s", scraped.Body.String(), line)
}
}
})
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestZoneKeyIsMaskedInTheVerdictsTheFailuresAndTheMetrics(t *testing.T) {
const (
key = "abcdefghijklmnopqrstuvwxyz"
keyed = key + ".xbl.dq.spamhaus.net"
masked = "********.xbl.dq.spamhaus.net"
)
synctest.Test(t, func(t *testing.T) {
var log bytes.Buffer
queue := newQueue()
p := dnsblParams(keyed)
p.Alerts = queue
p.ProcessLog = slog.New(slog.NewJSONHandler(&log, nil))
dnsbl := newDNSBL(&resolverStandIn{answers: map[string]answer{
"99.2.0.192." + keyed + ".": {addrs: []string{listing}},
"100.2.0.192." + keyed + ".": {rcode: serverFailure},
}}, p)
m := metrics.New(1, "app")
m.AddReputation(reputation.New(params()), dnsbl)
// Both clients are asked about before either answer comes, so that
// the failure does not keep the zone from the other query.
wantZones(t, dnsbl, listed)
wantZones(t, dnsbl, unlisted)
synctest.Wait()
wantZones(t, dnsbl, listed, masked)
if got := waiting(queue); len(got) != 1 || got[0].Detail["source"] != masked {
t.Errorf("alerts waiting %+v, want the failure's, from %s", got, masked)
}
scraped := httptest.NewRecorder()
m.ServeHTTP(scraped, httptest.NewRequestWithContext(t.Context(), http.MethodGet,
"/", http.NoBody))
for name, shown := range map[string]string{
"the log": log.String(), "the metrics": scraped.Body.String(),
} {
if strings.Contains(shown, key) || !strings.Contains(shown, masked) {
t.Errorf("%s shows the key, or does not name the zone:\n%s", name, shown)
}
}
})
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestVerdictsKeptAcrossARestart(t *testing.T) {
synctest.Test(t, func(t *testing.T) {
resolver := &resolverStandIn{answers: map[string]answer{
listedName: {addrs: []string{listing}},
}}
dnsbl := newDNSBL(resolver, dnsblParams(zone))
fetched := time.Now()
wantZones(t, dnsbl, unlisted)
wantZones(t, dnsbl, listed)
synctest.Wait()
kept := dnsbl.Snapshot()
want := []reputation.Verdict{
{Zone: zone, Client: netip.MustParseAddr(listed), Listed: true, Fetched: fetched},
{Zone: zone, Client: netip.MustParseAddr(unlisted), Fetched: fetched},
}
if !reflect.DeepEqual(kept, want) {
t.Errorf("verdicts %+v, want %+v", kept, want)
}
// Restarted an hour later with what reputation.json keeps, it uses
// the verdicts, and asks the zone nothing, until the TTL has passed
// since they were fetched.
time.Sleep(time.Hour)
restarted := &resolverStandIn{}
again := newDNSBL(restarted, dnsblParams(zone))
again.Load(kept)
wantZones(t, again, listed, zone)
wantZones(t, again, unlisted)
synctest.Wait()
wantAsked(t, restarted)
time.Sleep(cacheTTL - time.Hour)
wantZones(t, again, listed)
synctest.Wait()
wantAsked(t, restarted, listedName)
})
}
func TestNeitherAVerdictOfAZoneNotNamedNorOnePastItsTTLIsKept(t *testing.T) {
t.Parallel()
now := time.Date(2026, 10, 7, 0, 0, 0, 0, time.UTC)
p := dnsblParams(zone)
p.Now = func() time.Time { return now }
dnsbl := reputation.NewDNSBL(p)
// The last verdict still in use, one fetched a TTL ago, and one of a
// zone SWWAF_DNSBL_ZONES does not name.
inUse := reputation.Verdict{
Zone: zone, Client: netip.MustParseAddr(listed), Listed: true,
Fetched: now.Add(-cacheTTL + time.Nanosecond),
}
stale := reputation.Verdict{
Zone: zone, Client: netip.MustParseAddr(unlisted), Fetched: now.Add(-cacheTTL),
}
notNamed := reputation.Verdict{
Zone: otherZone, Client: netip.MustParseAddr(listed), Listed: true, Fetched: now,
}
dnsbl.Load([]reputation.Verdict{notNamed, stale, inUse})
if got := dnsbl.Snapshot(); !reflect.DeepEqual(got, []reputation.Verdict{inUse}) {
t.Errorf("verdicts %+v, want only %+v", got, inUse)
}
}
func TestAtMost100000VerdictsKeptTheOneFetchedLongestAgoDroppedFirst(t *testing.T) {
t.Parallel()
now := time.Date(2026, 10, 7, 0, 0, 0, 0, time.UTC)
p := dnsblParams(zone)
p.Now = func() time.Time { return now }
dnsbl := reputation.NewDNSBL(p)
// 100,001 verdicts, listed by client, as reputation.json lists them,
// each fetched a millisecond before the one before it: the last is one
// too many.
const count = 100001
verdicts := make([]reputation.Verdict, 0, count)
client := netip.MustParseAddr("198.18.0.0")
for i := range count {
verdicts = append(verdicts, reputation.Verdict{
Zone: zone, Client: client, Fetched: now.Add(-time.Duration(i) * time.Millisecond),
})
client = client.Next()
}
dnsbl.Load(verdicts)
got := dnsbl.Snapshot()
if len(got) != count-1 || !slices.Contains(got, verdicts[0]) ||
slices.Contains(got, verdicts[count-1]) {
t.Errorf("%d verdicts kept, want all but the one fetched longest ago", len(got))
}
}
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestQueriesGoToTheResolverSWWAFDNSBLResolverNames(t *testing.T) {
resolver := &resolverStandIn{answers: map[string]answer{
listedName: {addrs: []string{listing}},
}}
conn, err := (&net.ListenConfig{}).ListenPacket(t.Context(), "udp", "127.0.0.1:0")
if err != nil {
t.Fatalf("listen: %v", err)
}
served := make(chan struct{})
go func() {
resolver.serveUDP(conn)
close(served)
}()
t.Cleanup(func() {
_ = conn.Close()
<-served
})
p := dnsblParams(zone)
p.Resolver = netip.MustParseAddrPort(conn.LocalAddr().String())
// On the real clock: the stand-in answers at once, so only a test
// process held up for a whole minute would see the query fail.
p.Timeout = time.Minute
isListed, err := reputation.NewDNSBL(p).LookUp(zone, netip.MustParseAddr(listed))
if err != nil || !isListed {
t.Errorf("listed %t (%v), want true", isListed, err)
}
wantAsked(t, resolver, listedName)
}
// resolverStandIn is a stand-in for the resolver the zones are asked
// through. It answers each query by the name asked about, as answers
// gives, with no such name for a name answers does not give, and not at
// all while hanging. It notes each name asked about.
type resolverStandIn struct {
mu sync.Mutex
answers map[string]answer
hanging bool
names []string
}
// answer is how the stand-in answers a name: with an A record of each of
// addrs, or with the response code rcode, unless it is 0, for no error.
type answer struct {
addrs []string
rcode uint16
}
// What the stand-in reads of a query, and writes in its reply.
const (
// headerLength is the length of a DNS message's header, which the
// question follows: its id, its flags, and how many questions,
// answers and other records it holds, two bytes each.
headerLength = 12
// typeAndClass is the length of the type and the class that end a
// question, after its name.
typeAndClass = 4
// replyFlags mark a reply to a query that asked for recursion, which
// is available, with no error. The response code goes in their last
// four bits.
replyFlags = 0x8180
// maxMessage is the longest query read over UDP.
maxMessage = 1232
)
// set has the stand-in answer name with given.
func (s *resolverStandIn) set(name string, given answer) {
s.mu.Lock()
defer s.mu.Unlock()
s.answers[name] = given
}
// dial connects Go's resolver to the stand-in through an in-memory
// connection, on which it sends each query, and reads each reply, after
// its length, as over TCP.
func (s *resolverStandIn) dial(context.Context, string, string) (net.Conn, error) {
client, server := net.Pipe()
go s.serve(server)
return client, nil
}
// serve answers the queries that come on conn until the resolver closes
// it.
func (s *resolverStandIn) serve(conn net.Conn) {
defer func() {
_ = conn.Close()
}()
for {
var length [2]byte
_, err := io.ReadFull(conn, length[:])
if err != nil {
return
}
message := make([]byte, binary.BigEndian.Uint16(length[:]))
_, err = io.ReadFull(conn, message)
if err != nil {
return
}
reply, answered := s.reply(message)
if !answered {
continue // the resolver gives up, and closes conn
}
//nolint:gosec // a reply of a few dozen bytes
_, err = conn.Write(append(binary.BigEndian.AppendUint16(nil, uint16(len(reply))),
reply...))
if err != nil {
return
}
}
}
// serveUDP answers the queries that come on conn, each in a datagram, as
// a resolver does, until conn is closed.
func (s *resolverStandIn) serveUDP(conn net.PacketConn) {
message := make([]byte, maxMessage)
for {
n, from, err := conn.ReadFrom(message)
if err != nil {
return
}
reply, answered := s.reply(message[:n])
if answered {
_, _ = conn.WriteTo(reply, from)
}
}
}
// reply returns the stand-in's reply to message, a query, and false for
// none, while it hangs. It notes the name asked about.
func (s *resolverStandIn) reply(message []byte) ([]byte, bool) {
// The name is labels, each after its length, ended by a length of 0.
var labels []string
end := headerLength
for message[end] != 0 {
length := int(message[end])
labels = append(labels, string(message[end+1:end+1+length]))
end += 1 + length
}
end += 1 + typeAndClass
name := strings.Join(labels, ".") + "."
s.mu.Lock()
s.names = append(s.names, name)
given, found := s.answers[name]
hanging := s.hanging
s.mu.Unlock()
if hanging {
return nil, false
}
if !found {
given = answer{rcode: noSuchName}
}
// The query's id, the flags, one question, the answers, and no other
// records, then the question, as asked.
reply := slices.Clone(message[:2])
reply = binary.BigEndian.AppendUint16(reply, replyFlags|given.rcode)
reply = binary.BigEndian.AppendUint16(reply, 1)
//nolint:gosec // a handful of answers
reply = binary.BigEndian.AppendUint16(reply, uint16(len(given.addrs)))
reply = append(reply, 0, 0, 0, 0)
reply = append(reply, message[headerLength:end]...)
// An A record starts with the name asked about, by a pointer to it in
// the question, then its type, A, its class, IN, how long it may be
// kept, 60 seconds, and the length of its address, 4 bytes.
record := []byte{0xc0, headerLength, 0, 1, 0, 1, 0, 0, 0, 60, 0, 4}
for _, addr := range given.addrs {
reply = append(reply, record...)
reply = append(reply, netip.MustParseAddr(addr).AsSlice()...)
}
return reply, true
}
// dnsblParams returns the DNSBLParams of zones, with the tests' cache TTL
// and timeout, by the bubble's clock, with alerts to a queue that sends
// none.
func dnsblParams(zones ...string) reputation.DNSBLParams {
return reputation.DNSBLParams{
Zones: zones,
CacheTTL: cacheTTL,
Timeout: timeout,
Now: time.Now,
ProcessLog: slog.New(slog.DiscardHandler),
Alerts: newQueue(),
}
}
// waitForTheResolver waits, on the bubble's clock, an hour, until Go's
// resolver has given up on every stand-in that does not answer: it waits
// for a server as long as /etc/resolv.conf has it wait, a few seconds,
// even after the query was given up, and a bubble cannot end before it.
func waitForTheResolver() {
time.Sleep(time.Hour)
}
// newDNSBL returns the DNSBL of p, asking resolver.
func newDNSBL(resolver *resolverStandIn, p reputation.DNSBLParams) *reputation.DNSBL {
dnsbl := reputation.NewDNSBL(p)
dnsbl.SetDial(resolver.dial)
return dnsbl
}
// wantZones checks the zones whose verdict dnsbl says lists client, as a
// request from client finds them.
func wantZones(t *testing.T, dnsbl *reputation.DNSBL, client string, want ...string) {
t.Helper()
got := dnsbl.ListedBy(t.Context(), netip.MustParseAddr(client))
if !slices.Equal(got, want) {
t.Errorf("%s is listed by %v, want %v", client, got, want)
}
}
// wantQueries checks how many queries dnsbl made to zone, and how many of
// them failed.
func wantQueries(t *testing.T, dnsbl *reputation.DNSBL, queries, failures int) {
t.Helper()
if dnsbl.Queries(zone) != queries || dnsbl.Failures(zone) != failures {
t.Errorf("%d queries and %d failures, want %d and %d", dnsbl.Queries(zone),
dnsbl.Failures(zone), queries, failures)
}
}
// wantAsked checks the names the stand-in was asked about, in any order.
func wantAsked(t *testing.T, resolver *resolverStandIn, want ...string) {
t.Helper()
resolver.mu.Lock()
got := slices.Sorted(slices.Values(resolver.names))
resolver.mu.Unlock()
slices.Sort(want)
if !slices.Equal(got, want) {
t.Errorf("asked about %v, want %v", got, want)
}
}
internal/reputation/export_test.go
+19 -1
View File
@@ -1,9 +1,27 @@
package reputation
import "net/http"
import (
"context"
"net"
"net/http"
"net/netip"
)
// SetTransport has l's fetches go through transport instead of the
// network.
func (l *Lists) SetTransport(transport http.RoundTripper) {
l.httpClient.Transport = transport
}
// SetDial has d's queries go through dial instead of the network.
func (d *DNSBL) SetDial(
dial func(ctx context.Context, network, address string) (net.Conn, error),
) {
d.resolver = &net.Resolver{PreferGo: true, Dial: dial}
}
// LookUp asks zone about addr at once, as a query in the background does,
// and returns whether zone lists addr.
func (d *DNSBL) LookUp(zone string, addr netip.Addr) (bool, error) {
return d.lookUp(context.Background(), query{zone: zone, client: addr})
}
internal/reputation/reputation.go
+4 -2
View File
@@ -2,8 +2,10 @@
// blocklists of SWWAF_BLOCKLIST_URLS, and the file of AS:percent lines
// SWWAF_ASN_LIMIT_PERCENT_URL names. It keeps the last good copy of each,
// whole, comment lines included, which is used while a fetch fails, and
// when each was last tried, which the state package writes to
// reputation.json and reads from it, so that a restart keeps them too.
// when each was last tried. It also asks the DNSBL zones of
// SWWAF_DNSBL_ZONES about clients, and keeps their verdicts. The state
// package writes all of these to reputation.json and reads them from it,
// so that a restart keeps them too.
package reputation
import (
internal/requestlog/requestlog.go
+4 -3
View File
@@ -35,8 +35,8 @@ const (
// rule.
ActionRuleBlocked = "rule_blocked"
// ActionDenied is a request refused because its client is in
// SWWAF_DENY_NETS, or in a blocklist while SWWAF_BLOCKLIST_ACTION is
// deny.
// SWWAF_DENY_NETS, in a blocklist while SWWAF_BLOCKLIST_ACTION is deny,
// or listed by a DNSBL zone while SWWAF_REPUTATION_ACTION is deny.
ActionDenied = "denied"
// ActionCountryDenied is a request refused for its client's country.
ActionCountryDenied = "country_denied"
@@ -138,7 +138,8 @@ type Line struct {
// Counts names its count: minute, hour or day for a rate limit, and
// minute_bytes, hour_bytes or day_bytes for a byte limit.
LimitHit string `json:"limit_hit,omitempty"`
// Reputation are the URLs of the blocklists that list the client.
// Reputation are the URLs of the blocklists that list the client, then
// the DNSBL zones whose verdict lists it, their keys masked.
Reputation []string `json:"reputation,omitempty"`
// Offence is the offence the request was held as, OffenceLimit.
Offence string `json:"offence,omitempty"`
internal/smallwebwaf/smallwebwaf.go
+1
View File
@@ -201,6 +201,7 @@ func loadStateFiles(
Limiter: server.Limiter,
GeoJS: server.GeoJS,
Lists: server.Lists,
DNSBL: server.DNSBL,
Alerts: alertQueue,
Anomalies: server.Anomalies,
Now: now,
internal/state/state.go
+41 -8
View File
@@ -2,7 +2,8 @@
// SWWAF_STATE_DIR, as the "Persistent state" section of SPEC.md describes:
// bans.json holds the bans, clients.json each client's counters and
// history, lookups.json GeoJS's answers, reputation.json the last try and
// last good copy of each list fetched from a URL, and alerts.json the
// last good copy of each list fetched from a URL and the DNSBL zones'
// verdicts, and alerts.json the
// cooldowns, the hour under way, the alerts waiting for each destination
// and the anomaly counters. Load
// reads them at start, Watch takes in an admin's edit of one while
@@ -76,13 +77,14 @@ type Params struct {
// is (SWWAF_STATE_COUNTER_INTERVAL).
WriteDelay time.Duration
CounterInterval time.Duration
// Ledger, Limiter, GeoJS, Lists, Alerts and Anomalies hold the state.
// Alerts also receive a file_error alert for an edit set aside, and for
// a write that fails while smallwebwaf runs.
// Ledger, Limiter, GeoJS, Lists, DNSBL, Alerts and Anomalies hold the
// state. Alerts also receive a file_error alert for an edit set aside,
// and for a write that fails while smallwebwaf runs.
Ledger *bans.Ledger
Limiter *ratelimit.Limiter
GeoJS *lookup.GeoJS
Lists *reputation.Lists
DNSBL *reputation.DNSBL
Alerts *alerts.Queue
Anomalies *anomaly.Counters
// Now tells the time by which the counters' buckets run out, normally
@@ -145,8 +147,9 @@ type lookupsFile struct {
// reputationFile is reputation.json, indented for an admin to read and
// edit, so that each line of a list's copy is on a line of its own.
type reputationFile struct {
Version int `json:"version"`
Lists []reputation.List `json:"lists"`
Version int `json:"version"`
Lists []reputation.List `json:"lists"`
Verdicts []reputation.Verdict `json:"verdicts"`
}
// alertsFile is alerts.json, indented for an admin to read and edit.
@@ -432,6 +435,7 @@ func (f *Files) takeIn(name string, data []byte, edit bool) (int, error) {
return 0, fmt.Errorf("%s: %w", path, err)
}
f.params.DNSBL.Load(file.Verdicts)
entries = len(file.Lists)
case alertsJSON:
waiting, err := f.takeInAlerts(path, data)
@@ -567,6 +571,7 @@ func (f *Files) encode(name string) ([]byte, error) {
case reputationJSON:
return encodeIndented(reputationFile{
Version: version, Lists: f.params.Lists.Snapshot(),
Verdicts: f.params.DNSBL.Snapshot(),
})
default: // alerts.json
held := f.params.Alerts.Snapshot()
@@ -729,8 +734,12 @@ func (f *lookupsFile) check(data []byte) error {
// check refuses a list without its URL, which would name no list, or the
// time it was last tried, which would have it fetched at once, and a copy
// of it without the time it was fetched, or without its lines, which hold
// the list.
func (f *reputationFile) check([]byte) error {
// the list. It refuses a verdict without its zone or its client, which
// would be about no one, whether the zone lists the client, or the time
// it was fetched, which would drop it. A verdict's listed is false for a
// client the zone does not list, which Verdicts cannot tell from a
// missing one, so each listed is read again as written.
func (f *reputationFile) check(data []byte) error {
for i, kept := range f.Lists {
switch {
case kept.URL == "":
@@ -744,6 +753,30 @@ func (f *reputationFile) check([]byte) error {
}
}
var written struct {
Verdicts []struct {
Listed *bool `json:"listed"`
} `json:"verdicts"`
}
err := json.Unmarshal(data, &written)
if err != nil {
return err
}
for i, verdict := range f.Verdicts {
switch {
case verdict.Zone == "":
return fmt.Errorf("verdicts %w", missing(i, "zone"))
case !verdict.Client.IsValid():
return fmt.Errorf("verdicts %w", missing(i, "client"))
case written.Verdicts[i].Listed == nil:
return fmt.Errorf("verdicts %w", missing(i, "listed"))
case verdict.Fetched.IsZero():
return fmt.Errorf("verdicts %w", missing(i, "fetched"))
}
}
return nil
}
internal/state/state_test.go
+74 -8
View File
@@ -38,9 +38,11 @@ const (
lookupsJSON = "lookups.json"
reputationJSON = "reputation.json"
alertsJSON = "alerts.json"
// blocklistURL and torURL are the blocklists the tests' lists name.
// blocklistURL and torURL are the blocklists the tests' lists name, and
// dnsblZone the DNSBL zone of the tests' verdicts.
blocklistURL = "https://lists.example/drop.txt"
torURL = "https://lists.example/tor.txt"
dnsblZone = "dnsbl.example"
// The AS number and AS name the tests' clients are looked up in.
asn = "AS64496"
asName = "Example Net"
@@ -210,7 +212,8 @@ const filledAlertsJSON = `{
`
// filledReputationJSON is reputation.json holding the blocklists' last
// tries and the copy of one, with its comment line, as fill puts them in.
// tries and the copy of one, with its comment line, and two verdicts of a
// DNSBL zone, as fill puts them in.
const filledReputationJSON = `{
"version": 1,
"lists": [
@@ -228,6 +231,20 @@ const filledReputationJSON = `{
"url": "https://lists.example/tor.txt",
"tried": "2026-10-06T00:00:00Z"
}
],
"verdicts": [
{
"zone": "dnsbl.example",
"client": "203.0.113.9",
"listed": true,
"fetched": "2026-10-05T23:00:00Z"
},
{
"zone": "dnsbl.example",
"client": "2001:db8::1",
"listed": false,
"fetched": "2026-10-05T22:00:00Z"
}
]
}
`
@@ -263,6 +280,8 @@ func TestFilesWrittenAndReadBack(t *testing.T) {
t.Errorf("%s read back\n%+v\nwant\n%+v", reputationJSON, got, want)
}
wantEqual(t, reputationJSON, after.DNSBL.Snapshot(), before.DNSBL.Snapshot())
if got, want := after.Alerts.Snapshot(), before.Alerts.Snapshot(); !reflect.DeepEqual(
got, want) {
t.Errorf("%s read back\n%+v\nwant\n%+v", alertsJSON, got, want)
@@ -422,8 +441,8 @@ func TestMissingFilesAreEmptyState(t *testing.T) {
held := params.Alerts.Snapshot()
if len(params.Ledger.Snapshot()) != 0 || len(params.Limiter.Snapshot()) != 0 ||
len(params.GeoJS.Snapshot()) != 0 || len(params.Lists.Snapshot()) != 0 ||
len(held.Cooldowns) != 0 || len(held.Waiting[alerts.DestinationWebhook]) != 0 ||
held.Hour.Sent != 0 {
len(params.DNSBL.Snapshot()) != 0 || len(held.Cooldowns) != 0 ||
len(held.Waiting[alerts.DestinationWebhook]) != 0 || held.Hour.Sent != 0 {
t.Error("state from no files")
}
}
@@ -592,6 +611,11 @@ func TestReputationJSONEntryWithoutAFieldItNeedsStopsTheStart(t *testing.T) {
drop = `"url": "` + blocklistURL + `", `
tried = `"tried": "2026-10-06T00:00:00Z", `
fetched = `"fetched": "2026-10-06T00:00:00Z"`
// verdictZone, verdictClient and listed start a verdict, which
// fetched ends.
verdictZone = `"zone": "` + dnsblZone + `", `
verdictClient = `"client": "198.51.100.7", `
listed = `"listed": false, `
)
for _, tc := range []struct {
@@ -621,6 +645,30 @@ func TestReputationJSONEntryWithoutAFieldItNeedsStopsTheStart(t *testing.T) {
`{"url": "` + torURL + `", ` + tried + fetched + `}]}`,
`: entry 2 has no "lines"`,
},
{
"a verdict without its zone",
`{"version": 1, "verdicts": [{` + verdictClient + listed + fetched + `}]}`,
`: verdicts entry 1 has no "zone"`,
},
{
"a verdict without its client",
`{"version": 1, "verdicts": [{` + verdictZone + listed + fetched + `}]}`,
`: verdicts entry 1 has no "client"`,
},
{
// A client the zone does not list has a listed of false, which is
// not having none.
"a verdict without whether the zone lists the client",
`{"version": 1, "verdicts": [{` + verdictZone + verdictClient + listed +
fetched + `}, {` + verdictZone + verdictClient + fetched + `}]}`,
`: verdicts entry 2 has no "listed"`,
},
{
"a verdict without the time it was fetched",
`{"version": 1, "verdicts": [{` + verdictZone + verdictClient +
`"listed": true}]}`,
`: verdicts entry 1 has no "fetched"`,
},
} {
t.Run(tc.name, func(t *testing.T) {
t.Parallel()
@@ -1133,7 +1181,8 @@ func TestEditOfEachFileTakenIn(t *testing.T) {
edit(t, dir, reputationJSON, `{"version": 1, "lists": [{"url": "`+blocklistURL+`", `+
`"tried": "2026-10-06T00:00:00Z", "fetched": "2026-10-06T00:00:00Z", `+
`"lines": ["198.51.100.7"]}]}`)
`"lines": ["198.51.100.7"]}], "verdicts": [{"zone": "`+dnsblZone+`", `+
`"client": "198.51.100.7", "listed": true, "fetched": "2026-10-06T00:00:00Z"}]}`)
wantTakenIn(t, lines, dir, reputationJSON)
listedBy := params.Lists.ListedBy(client.Addr())
@@ -1142,6 +1191,10 @@ func TestEditOfEachFileTakenIn(t *testing.T) {
client.Addr(), listedBy)
}
wantEqual(t, reputationJSON, params.DNSBL.Snapshot(), []reputation.Verdict{{
Zone: dnsblZone, Client: client.Addr(), Listed: true, Fetched: midnight(),
}})
// A netblock with bits past its length is read as the netblock it is
// in.
edit(t, dir, alertsJSON, `{"version": 1, "cooldowns": [{"event": "ban", `+
@@ -1558,6 +1611,10 @@ func newParams(dir string) state.Params {
BlocklistURLs: []string{blocklistURL, torURL}, Refresh: 24 * time.Hour,
Now: midnight, ProcessLog: discard, Alerts: queue,
}),
DNSBL: reputation.NewDNSBL(reputation.DNSBLParams{
Zones: []string{dnsblZone}, CacheTTL: 24 * time.Hour, Timeout: time.Second,
Now: midnight, ProcessLog: discard, Alerts: queue,
}),
Alerts: queue,
Anomalies: anomaly.New(anomaly.Params{
Net: anomaly.Thresholds{RequestsPerMinute: 1000},
@@ -1582,9 +1639,10 @@ func office() netip.Prefix {
// fill puts a permanent ban an admin made, a ban for a broken limit and
// one for a clear sign of attack, clients with counts and histories,
// GeoJS answers, the blocklists' last tries and the copy of one, as
// filledReputationJSON holds them, and alerts and anomaly counters, as
// filledAlertsJSON holds them, into the parts of params.
// GeoJS answers, the blocklists' last tries and the copy of one, and two
// verdicts of a DNSBL zone, as filledReputationJSON holds them, and alerts
// and anomaly counters, as filledAlertsJSON holds them, into the parts of
// params.
func fill(params state.Params) {
now := midnight()
client := netip.MustParsePrefix("203.0.113.9/32")
@@ -1631,6 +1689,14 @@ func fill(params state.Params) {
panic(err) // the copy reads
}
params.DNSBL.Load([]reputation.Verdict{
{
Zone: dnsblZone, Client: netip.MustParseAddr("2001:db8::1"),
Fetched: now.Add(-2 * time.Hour),
},
{Zone: dnsblZone, Client: client.Addr(), Listed: true, Fetched: now.Add(-time.Hour)},
})
// An alert waiting, a repeat of it the cooldown holds back, another
// alert waiting, and one past the two an hour, for the hour's summary.
ban := alerts.Alert{