1 Commits
Author SHA1 Message Date
sneak a8b978dfc9 Name request log actions so they say what they mean (closes #229)
check / check (push) Successful in 4m12s
Rename too_large, rate_limited, rule_blocked and denied, and split banned
into temporarily_banned and permanently_banned by whether the ban runs out.
A client under both a temporary and a permanent ban gets the permanent one.
A request that makes a ban permanent gets permanently_banned. A ban rule, a
trap path or a CrowdSec decision gets the value for the ban it makes, and in
observe mode would_action names the ban enforce mode would have made.

The offence kind rule_blocked follows, as a metric label and a key in
clients.json.

Model: opus-5-5
2026-11-09 04:34:40 +00:00
14 changed files with 17 additions and 1043 deletions

No files matched your search

+3 -6
View File
@@ -265,7 +265,6 @@ Every setting the code reads, with its default:
| `SWW_TAG_LISTS` | empty | Address lists that each add a tag to their clients' requests, as `name=URL` pairs, such as `torexit=https://example.com/tor.txt`. |
| `SWW_TAG_DENY` | empty | Tags whose requests are refused, such as `torexit,country=CN`. |
| `SWW_TAG_LIMIT_PERCENT` | empty | Tags whose clients get a percentage, from 1 to 99, of every request and byte limit, such as `vpn:50,asn=14061:25`; the lowest applies. |
| `SWW_ROUTEWATCH_URL` | unset | The URL of a routewatch instance, which gives each client's `asn=` and `prefix=` tags. Unset, nothing asks routewatch. |
| `SWW_ASN_BYTES_PERCENT` | empty | A percentage of the byte limits alone for the AS numbers listed, in place of their `asn=` tag's percentage. |
| `SWW_COUNTRY_BYTES_PERCENT` | empty | A percentage of the byte limits alone for the countries listed, in place of their `country=` tag's percentage. |
| `SWW_UNKNOWN_LIMIT_PERCENT` | `100` | The percentage of every limit a client without a country gets. |
@@ -568,9 +567,7 @@ the `tags` field of its log line:
or `vpn=mullvad`;
- `crowdsec`, `dnsbl=<zone>` and `abuseipdb`, when that reputation source lists
the client;
- `exploitsignature`, when a rule file or the Core Rule Set matched the request;
- `asn=<number>` and `prefix=<netblock>`, from a routewatch instance, while
`SWW_ROUTEWATCH_URL` is set.
- `exploitsignature`, when a rule file or the Core Rule Set matched the request.
`SWW_TAG_DENY` refuses requests by tag, and `SWW_TAG_LIMIT_PERCENT` lowers a
client's limits by tag. See "Request tags" in [`SPEC.md`](SPEC.md).
@@ -633,8 +630,8 @@ is set aside as `<name>.bad`, and never stops the running process (see
- `internal/discriminator`: `discriminator`'s server, which runs the checks and
answers its own endpoints.
- `internal/ratelimit`, `internal/bans`, `internal/rules`, `internal/waf`,
`internal/lookup`, `internal/reputation`, `internal/routewatch`,
`internal/anomaly` and `internal/alerts`: one check or source each.
`internal/lookup`, `internal/reputation`, `internal/anomaly` and
`internal/alerts`: one check or source each.
- `internal/state`, `internal/requestlog`, `internal/remotelog` and
`internal/metrics`: the state files, the log lines, the syslog sender and the
metrics.
+2 -23
View File
@@ -87,8 +87,6 @@ to them.
database.
- `internal/reputation`: fetches the lists, asks the DNS blocklists and
AbuseIPDB in the background, and keeps their answers.
- `internal/routewatch`: asks a routewatch instance about clients in the
background, and keeps its answers.
- `internal/anomaly`: the anomaly counters and their alerts.
- `internal/alerts`: holds back repeats and alerts past the hourly limit,
and sends the rest to the webhook, Slack and ntfy, each from a queue of
@@ -146,9 +144,7 @@ request log's `heuristic` field gives that name. `refusal_limit` and
around its address, a /64 by default, since one abuser usually controls a
whole /64. A client is therefore one IPv4 address or one IPv6 group.
3. The tag lists (`tag_lists`; see "Request tags"). Each list that holds the
client adds its tag. Next, `routewatch_tags` adds the client's `asn=` and
`prefix=` tags while `SWW_ROUTEWATCH_URL` is set. A client in
`SWW_ALLOW_NETS` gets no tag.
client adds its tag. A client in `SWW_ALLOW_NETS` gets no tag.
4. The static lists. A client in `SWW_ALLOW_NETS` skips steps 5 to 15, the byte
limits and the error burst. A client in `SWW_DENY_NETS` is refused
(`deny_nets`); a client in both lists is let through.
@@ -565,21 +561,6 @@ Each tag and where it comes from:
- `exploitsignature`, when a rule file or the Core Rule Set matched the request.
For the Core Rule Set a match is a score at or over
`SWW_WAF_ANOMALY_THRESHOLD`.
- `asn=<number>` and `prefix=<netblock>` from a routewatch instance
(`routewatch_tags`), while `SWW_ROUTEWATCH_URL` is set:
- The tag `prefix=` holds the netblock the client's address is announced in,
such as `prefix=192.0.2.0/24`.
- An `asn=` tag has the same spelling as the lookup's. When routewatch and
the lookup give the same AS number, a request has one `asn=` tag. When
they differ, it has both.
- They are added before `deny_nets`, so the log line of a request it refused
has them.
- `discriminator` asks routewatch in the background and never waits for it.
An answer is kept for 1 hour, in memory only.
- After a query fails, routewatch is not asked for 1 minute.
- A request without these tags gives the reason in the log field
`routewatch_tags_missing` (see
[`docs/request-log.md`](docs/request-log.md)).
A client in `SWW_ALLOW_NETS` skips the checks and gets no tag. Every tag is in
lower case, except a country code, which is in capitals. `SWW_TAG_DENY` and
@@ -598,9 +579,7 @@ Two settings act on tags:
The tag settings are empty by default, for the same reason as the country list.
The lookup waits for GeoJS when either names an `asn=` or `country=` tag. A
setting that names one stops the start while `SWW_LOOKUP_SOURCE` is `off`. The
exception is an `asn=` tag while `SWW_ROUTEWATCH_URL` is set, since routewatch
gives that tag.
setting that names one stops the start while `SWW_LOOKUP_SOURCE` is `off`.
### Biased thresholds
-6
View File
@@ -175,12 +175,6 @@ Every one of these lines has:
- `abuseipdb`.
- `tags`: the request's tags, sorted, left out when it has none. Each tag and
where it comes from is under "Request tags" in [`SPEC.md`](../SPEC.md).
- `routewatch_tags_missing`: why the request got no tags from routewatch,
present only while `SWW_ROUTEWATCH_URL` is set:
- `no_answer_yet`: routewatch has not answered about this client yet;
- `request_failed`: a query to routewatch failed within the last minute, and
the process log says why;
- `no_route`: routewatch knows no live route for the address.
- `offence`: `limit`, for a request that broke a limit.
- `ban_expires`: when the client's ban ends, for a request that made one or was
refused under one: a time, or `permanent`.
+2 -11
View File
@@ -145,8 +145,8 @@ cut off; the client is banned.
the start while a setting that needs a lookup is set:
- `SWW_EXCLUSIVELY_ALLOWED_COUNTRIES`;
- `SWW_ADD_LOOKUP_HEADERS`;
- `SWW_TAG_DENY` or `SWW_TAG_LIMIT_PERCENT` naming a `country=` tag, or an
`asn=` tag while `SWW_ROUTEWATCH_URL` is unset;
- `SWW_TAG_DENY` or `SWW_TAG_LIMIT_PERCENT` naming an `asn=` or `country=`
tag;
- the two byte percentage lists and `SWW_ASN_LIMIT_PERCENT_URL`;
- `SWW_UNKNOWN_LIMIT_PERCENT` below `100`;
- an anomaly threshold per AS number.
@@ -186,20 +186,11 @@ Every request gets tags from what `discriminator` knows of its client. See
`exploitsignature`: a request gets that tag after the rate limits have counted
it.
- `SWW_ROUTEWATCH_URL` (default unset): the `http` or `https` URL of a
routewatch instance. [routewatch](https://git.eeqj.de/sneak/routewatch) is a
service that gives an address's announced netblock and origin AS.
`discriminator` asks it about each client, in the background, and adds the
tags `asn=<number>` and `prefix=<netblock>` from its answer. While it is
unset, nothing asks routewatch and no request gets a `prefix=` tag.
`SWW_TAG_DENY` and `SWW_TAG_LIMIT_PERCENT` accept tags in either case. A tag
that no request can get stops the start. A request can get only these:
- a name in `SWW_TAG_LISTS`;
- an `asn=` or `country=` tag;
- a `prefix=` tag with a netblock, while `SWW_ROUTEWATCH_URL` is set. It matches
only that exact announced netblock, such as `prefix=192.0.2.0/24`;
- a `dnsbl=` tag of a zone in `SWW_DNSBL_ZONES`, in lower case, with a key
masked as the `reputation` field shows it;
- `abuseipdb`, while `SWW_ABUSEIPDB_KEY` is set;
+9 -58
View File
@@ -12,7 +12,6 @@ import (
"errors"
"fmt"
"log/slog"
"maps"
"math"
"net"
"net/http"
@@ -107,9 +106,6 @@ type Config struct {
TagLists []TagList
TagDeny []string
TagLimitPercent map[string]int64
// RouteWatchURL is where the routewatch instance is that adds the asn=
// and prefix= tags, "" for none (SWW_ROUTEWATCH_URL).
RouteWatchURL string
// The biased thresholds. ASNBytesPercent and CountryBytesPercent give
// those they list a percentage of the byte limits in place of the one
// TagLimitPercent gives their asn= or country= tag
@@ -466,9 +462,6 @@ var (
errNotLAPIURL = errors.New(
"is not an http or https URL without a user or a fragment, " +
"such as http://172.17.0.1:8080")
errNotRouteWatchURL = errors.New(
"is not an http or https URL without a user or a fragment, " +
"such as https://routewatch.example")
errNeedsLAPIKey = errors.New("the engine answers no request without it")
errLAPIKeyUnused = errors.New("it is sent only to the engine at that URL")
errAnotherList = errors.New(
@@ -482,9 +475,7 @@ var (
"is not a tag, : and a percentage, such as vpn:50 or asn=14061:25")
errNotTagPercent = errors.New("is not a percentage from 1 to 99")
errNotASNTag = errors.New("is not asn= and a number, such as asn=14061")
errNotPrefixTag = errors.New(
"is not prefix= and a netblock, such as prefix=192.0.2.0/24")
errExploitTag = errors.New(
errExploitTag = errors.New(
"cannot lower limits: the rate limits count a request before its " +
"rules are matched")
errCrowdSecTag = errors.New(
@@ -492,9 +483,8 @@ var (
errNoRequestGetsTag = errors.New(
"is a tag no request can get: it is not a name in SWW_TAG_LISTS, an " +
"asn= or country= tag, a dnsbl= tag of a zone in SWW_DNSBL_ZONES, " +
"a prefix= tag with SWW_ROUTEWATCH_URL set, abuseipdb with " +
"SWW_ABUSEIPDB_KEY set, or exploitsignature with rule files or the " +
"Core Rule Set on")
"abuseipdb with SWW_ABUSEIPDB_KEY set, or exploitsignature with " +
"rule files or the Core Rule Set on")
)
// FromEnvironment reads the settings with lookupEnv, normally
@@ -534,7 +524,6 @@ func FromEnvironment(
TagLists: env.tagLists("SWW_TAG_LISTS"),
TagDeny: env.tags("SWW_TAG_DENY"),
TagLimitPercent: env.tagPercents("SWW_TAG_LIMIT_PERCENT"),
RouteWatchURL: env.routeWatchURL("SWW_ROUTEWATCH_URL"),
ASNBytesPercent: env.percents("SWW_ASN_BYTES_PERCENT", ParseASN),
CountryBytesPercent: env.percents("SWW_COUNTRY_BYTES_PERCENT", parseCountry),
UnknownLimitPercent: env.percent("SWW_UNKNOWN_LIMIT_PERCENT", "100"),
@@ -1030,22 +1019,6 @@ func (e *environment) crowdSecDecisionsURL(name string) string {
return lapi.JoinPath("v1", "decisions").String()
}
// routeWatchURL reads the setting that is the URL of a routewatch instance,
// "" while it is unset or empty.
func (e *environment) routeWatchURL(name string) string {
value := e.value(name, "")
if value == "" {
return ""
}
instance, err := url.Parse(value)
if err != nil || !isHTTPURL(instance) {
e.check(name, fmt.Errorf("%q %w", value, errNotRouteWatchURL))
}
return value
}
// lookupSource reads the setting that is where clients are looked up:
// geojs, file, or off.
func (e *environment) lookupSource(name, defaultValue string) string {
@@ -1091,21 +1064,14 @@ func (e *environment) checkCountriesAndLookups(cfg *Config) {
return
}
// routewatch gives asn= tags too.
onlyLookupGives := func(tag string) bool {
return isLookupTag(tag) &&
(cfg.RouteWatchURL == "" || !strings.HasPrefix(tag, ASNTagPrefix))
}
for _, setting := range []struct {
name string
set bool
}{
{"SWW_EXCLUSIVELY_ALLOWED_COUNTRIES", len(cfg.ExclusivelyAllowedCountries) > 0},
{"SWW_ADD_LOOKUP_HEADERS", cfg.AddLookupHeaders},
{"SWW_TAG_DENY", slices.ContainsFunc(cfg.TagDeny, onlyLookupGives)},
{"SWW_TAG_LIMIT_PERCENT", slices.ContainsFunc(
slices.Collect(maps.Keys(cfg.TagLimitPercent)), onlyLookupGives)},
{"SWW_TAG_DENY", slices.ContainsFunc(cfg.TagDeny, isLookupTag)},
{"SWW_TAG_LIMIT_PERCENT", tagLimitNeedsLookups(cfg)},
{"SWW_ASN_BYTES_PERCENT", len(cfg.ASNBytesPercent) > 0},
{"SWW_COUNTRY_BYTES_PERCENT", len(cfg.CountryBytesPercent) > 0},
{"SWW_UNKNOWN_LIMIT_PERCENT", cfg.UnknownLimitPercent < 100},
@@ -1205,10 +1171,6 @@ func requestCanGetTag(cfg *Config, tag string) bool {
return cfg.AbuseIPDBKey != ""
}
if strings.HasPrefix(tag, PrefixTagPrefix) {
return cfg.RouteWatchURL != ""
}
zone, isZoneTag := strings.CutPrefix(tag, DNSBLTagPrefix)
return isZoneTag && slices.ContainsFunc(cfg.DNSBLZones, func(each string) bool {
@@ -1217,7 +1179,7 @@ func requestCanGetTag(cfg *Config, tag string) bool {
}
// TagsNeedLookup reports whether SWW_TAG_DENY or SWW_TAG_LIMIT_PERCENT
// names an asn= or country= tag.
// names an asn= or country= tag, which only the lookup can give a request.
func (c *Config) TagsNeedLookup() bool {
return slices.ContainsFunc(c.TagDeny, isLookupTag) || tagLimitNeedsLookups(c)
}
@@ -1961,7 +1923,6 @@ const (
ASNTagPrefix = "asn="
CountryTagPrefix = "country="
DNSBLTagPrefix = "dnsbl="
PrefixTagPrefix = "prefix="
AbuseIPDBTag = "abuseipdb"
CrowdSecTag = "crowdsec"
// ExploitSignatureTag is the tag of a request a rule file or the Core
@@ -2038,10 +1999,9 @@ func parseTags(value string) ([]string, error) {
// parseTag reads a tag in either case, with spaces around it, and returns it
// as a request carries it: in lower case, but for a country code, which is in
// capitals, an asn= tag's number without leading zeros, and a prefix= tag's
// netblock as routewatch writes it. An asn=, country= or prefix= tag must name
// a number, a country or a netblock, so that a mistyped one stops the start
// instead of never matching.
// capitals, and an asn= tag's number without leading zeros. An asn= or
// country= tag must name a number or a country, so that a mistyped one stops
// the start instead of never matching.
func parseTag(value string) (string, error) {
tag := strings.ToLower(strings.TrimSpace(value))
@@ -2063,15 +2023,6 @@ func parseTag(value string) (string, error) {
return CountryTagPrefix + country, nil
}
if text, isPrefix := strings.CutPrefix(tag, PrefixTagPrefix); isPrefix {
netblock, err := netip.ParsePrefix(text)
if err != nil || netblock != netblock.Masked() {
return "", fmt.Errorf("%q %w", value, errNotPrefixTag)
}
return PrefixTagPrefix + netblock.String(), nil
}
return tag, nil
}
-59
View File
@@ -61,7 +61,6 @@ const (
reputationAction = "SWW_REPUTATION_ACTION"
reputationCacheTTL = "SWW_REPUTATION_CACHE_TTL"
reputationTimeout = "SWW_REPUTATION_TIMEOUT"
routeWatchURL = "SWW_ROUTEWATCH_URL"
crowdSecURL = "SWW_CROWDSEC_LAPI_URL"
crowdSecKey = "SWW_CROWDSEC_LAPI_KEY"
banResponse = "SWW_BAN_RESPONSE"
@@ -1404,63 +1403,6 @@ func TestTagSettingsAsSet(t *testing.T) {
}
}
func TestRouteWatchURLAndPrefixTags(t *testing.T) {
t.Parallel()
const rw = "https://routewatch.example"
if cfg := fromEnvironment(t, environment{}); cfg.RouteWatchURL != "" {
t.Errorf("default %s is %q, want unset", routeWatchURL, cfg.RouteWatchURL)
}
cfg := fromEnvironment(t, environment{
routeWatchURL: rw, tagDeny: "prefix=2001:DB8::/32",
tagLimitPercent: "prefix=192.0.2.0/24:50",
})
wantDeny := []string{"prefix=2001:db8::/32"}
if cfg.RouteWatchURL != rw || !slices.Equal(cfg.TagDeny, wantDeny) ||
cfg.TagLimitPercent["prefix=192.0.2.0/24"] != 50 {
t.Errorf("gave %q, %v, %v", cfg.RouteWatchURL, cfg.TagDeny, cfg.TagLimitPercent)
}
// routewatch gives asn= tags too, so an asn= tag needs no lookups while it is set.
fromEnvironment(t, environment{
lookupSource: off, routeWatchURL: rw, tagDeny: "asn=64496",
})
for _, stopped := range []struct {
env environment
name string
}{
{environment{routeWatchURL: "ftp://routewatch.example"}, routeWatchURL},
{environment{routeWatchURL: "routewatch.example"}, routeWatchURL},
{environment{tagDeny: "prefix=192.0.2.0/24"}, tagDeny},
{environment{routeWatchURL: rw, tagDeny: "prefix=nowhere"}, tagDeny},
{environment{routeWatchURL: rw, tagDeny: "prefix=192.0.2.5/24"}, tagDeny},
{
environment{routeWatchURL: rw, tagLimitPercent: "prefix=192.0.2.0/33:50"},
tagLimitPercent,
},
{environment{lookupSource: off, routeWatchURL: rw, tagDeny: "country=DE"}, tagDeny},
{environment{lookupSource: off, tagDeny: "asn=64496"}, tagDeny},
} {
_, err := config.FromEnvironment(stopped.env.lookupEnv)
if err == nil || !strings.HasPrefix(err.Error(), stopped.name+": ") {
t.Errorf("%v gave the error %v, want one naming %s", stopped.env, err, stopped.name)
}
}
badURL := environment{routeWatchURL: "ftp://routewatch.example"}
_, err := config.FromEnvironment(badURL.lookupEnv)
want := routeWatchURL + `: "ftp://routewatch.example" is not an http or https URL ` +
"without a user or a fragment, such as https://routewatch.example"
if err == nil || err.Error() != want {
t.Errorf("error %v, want %s", err, want)
}
}
func TestTagThatNoRequestCanGetIsToldApartFromOneItCan(t *testing.T) {
t.Parallel()
@@ -2561,7 +2503,6 @@ func TestLogsEachSettingWithItsValue(t *testing.T) {
reputationAction: "limit:25",
reputationCacheTTL: defaultReputationCacheTTL,
reputationTimeout: "2s",
routeWatchURL: "",
crowdSecURL: "",
crowdSecKey: "",
banResponse: "403",
-1
View File
@@ -123,7 +123,6 @@ func New(params Params) *Server {
Lists: h.lists,
DNSBL: h.dnsbl,
AbuseIPDB: h.abuseIPDB,
RouteWatch: h.routeWatch,
Metrics: h.metrics,
}
}
-1
View File
@@ -44,7 +44,6 @@ const (
func beforeLookup() []heuristic {
return []heuristic{
tagLists{},
routeWatchTags{},
denyNets{},
activeBan{},
}
+1 -11
View File
@@ -20,7 +20,6 @@ import (
"sneak.berlin/go/smallwebwaf/internal/metrics"
"sneak.berlin/go/smallwebwaf/internal/ratelimit"
"sneak.berlin/go/smallwebwaf/internal/reputation"
"sneak.berlin/go/smallwebwaf/internal/routewatch"
"sneak.berlin/go/smallwebwaf/internal/rules"
"sneak.berlin/go/smallwebwaf/internal/waf"
)
@@ -86,8 +85,7 @@ type Params struct {
// Server is the server discriminator runs, with the parts whose state the state
// files keep, the lookup database, nil unless SWW_LOOKUP_SOURCE is file, the
// lists fetched from URLs, which its Run fetches, the DNSBL zones' verdicts,
// AbuseIPDB's scores and checks spent, and the metrics. RouteWatch, nil while
// SWW_ROUTEWATCH_URL is unset, keeps its answers in memory only.
// AbuseIPDB's scores and checks spent, and the metrics.
type Server struct {
*http.Server
@@ -99,7 +97,6 @@ type Server struct {
Lists *reputation.Lists
DNSBL *reputation.DNSBL
AbuseIPDB *reputation.AbuseIPDB
RouteWatch *routewatch.Client
Metrics *metrics.Metrics
}
@@ -155,12 +152,6 @@ func newHandler(params Params) *handler {
coreRuleSet: params.CoreRuleSet,
alerts: params.Alerts,
}
if params.Config.RouteWatchURL != "" {
h.routeWatch = routewatch.New(routewatch.Params{
URL: params.Config.RouteWatchURL, Now: params.Now, ProcessLog: params.ProcessLog,
})
}
h.geojs = lookup.New(lookup.Params{
URL: params.GeoJSURL,
Timeout: params.Config.LookupTimeout,
@@ -264,7 +255,6 @@ type handler struct {
lists *reputation.Lists
dnsbl *reputation.DNSBL
abuseIPDB *reputation.AbuseIPDB
routeWatch *routewatch.Client
rules *rules.Files
coreRuleSet *waf.CoreRuleSet
alerts *alerts.Queue
-42
View File
@@ -1,42 +0,0 @@
package discriminator
// The heuristic routewatch_tags. It reads SWW_ROUTEWATCH_URL.
import (
"context"
"strconv"
"time"
"sneak.berlin/go/smallwebwaf/internal/config"
)
// routeWatchTags adds to the request the tags asn=<number> and prefix=<netblock>
// from what routewatch answered about the client. It refuses nothing, and the
// request does not wait for routewatch. A request that gets no tags from it has
// the reason in the log line's routewatch_tags_missing. While SWW_ROUTEWATCH_URL
// is unset it does nothing. Answers are kept by client and asked for by the
// request's address, and a route without an origin AS gives only the prefix= tag.
type routeWatchTags struct{}
func (routeWatchTags) name() string { return "routewatch_tags" }
func (routeWatchTags) check(ctx context.Context, rq *request, _ time.Time) *verdict {
if rq.h.routeWatch == nil {
return nil
}
route, missing := rq.h.routeWatch.Route(ctx, rq.h.clientGroup(rq.client), rq.client)
if missing != "" {
rq.line.RouteWatchTagsMissing = string(missing)
return nil
}
rq.addTag(config.PrefixTagPrefix + route.Netblock.String())
if route.ASN != 0 {
rq.addTag(config.ASNTagPrefix + strconv.FormatUint(uint64(route.ASN), 10))
}
return nil
}
@@ -1,245 +0,0 @@
package discriminator_test
import (
"context"
"encoding/json"
"net/http"
"net/http/httptest"
"net/netip"
"strings"
"testing"
"time"
"sneak.berlin/go/smallwebwaf/internal/discriminator"
"sneak.berlin/go/smallwebwaf/internal/requestlog"
)
// routeWatchURL is SWW_ROUTEWATCH_URL, and tagsMissing the log line's field
// that says why a request got no tags from routewatch.
const (
routeWatchURL = "SWW_ROUTEWATCH_URL"
tagsMissing = "routewatch_tags_missing"
fromV6 = "2001:db8:1::5"
prefixV6 = "2001:db8::/32"
prefixDE = "203.0.113.0/24"
prefixKP = "198.51.100.0/24"
)
// standInRoute is what the stand-in for routewatch answers about an address.
type standInRoute struct {
Netblock string `json:"netblock"`
ASN int `json:"asn"`
}
// startRouteWatch starts a stand-in for routewatch's IP info API that knows the
// routes given by address, and returns its URL.
func startRouteWatch(t *testing.T, routes map[string]standInRoute) string {
t.Helper()
server := httptest.NewServer(http.HandlerFunc(
func(w http.ResponseWriter, r *http.Request) {
address := strings.TrimPrefix(r.URL.Path, "/api/v1/ip/")
result := map[string]any{"ip": address}
if route, known := routes[address]; known {
result["netblock"], result["asn"] = route.Netblock, route.ASN
}
_ = json.NewEncoder(w).Encode(map[string]any{
"status": "ok",
"data": map[string]any{"results": []any{result}},
})
}))
t.Cleanup(server.Close)
return server.URL
}
// waitForRouteWatch waits until routewatch's answer about from, in the client
// group of the given length, is kept, so that the next request gets its tags.
func waitForRouteWatch(
t *testing.T, server *discriminator.Server, from string, bits int,
) {
t.Helper()
addr := netip.MustParseAddr(from)
client := netip.PrefixFrom(addr, bits).Masked()
deadline := time.Now().Add(5 * time.Second)
for {
_, missing := server.RouteWatch.Route(context.Background(), client, addr)
if missing == "" {
return
}
if time.Now().After(deadline) {
t.Fatalf("routewatch's answer about %s not kept: %q", from, missing)
}
time.Sleep(time.Millisecond)
}
}
func wantTagsMissing(t *testing.T, line logLine, want string) {
t.Helper()
got, _ := line.fields[tagsMissing].(string)
if got != want {
t.Errorf("log line has %s %q, want %q", tagsMissing, got, want)
}
}
func TestRouteWatchTagsAreOnTheLogLineWithTheLookupsTags(t *testing.T) {
t.Parallel()
url := startRouteWatch(t, map[string]standInRoute{fromDE: {prefixDE, 64496}})
s, server, _ := startWithLookups(t, map[string]string{routeWatchURL: url})
waitForRouteWatch(t, server, fromDE, 32)
line := s.get(fromDE, http.StatusOK, requestlog.ActionForward)
wantTags(t, line, "asn=64496", "country=DE", "prefix="+prefixDE)
wantTagsMissing(t, line, "")
}
func TestRouteWatchTagsAreOnALineDenyNetsRefused(t *testing.T) {
t.Parallel()
url := startRouteWatch(t, map[string]standInRoute{fromKP: {prefixKP, 64511}})
s, _, server := startWithClock(t, "", map[string]string{
lookupSource: off, routeWatchURL: url, denyNets: fromKP,
})
waitForRouteWatch(t, server, fromKP, 32)
line := s.get(fromKP, http.StatusForbidden, requestlog.ActionClientDenied)
wantHeuristic(t, line, "deny_nets", `{"netblock": "`+fromKP+`/32"}`)
wantTags(t, line, "asn=64511", "prefix="+prefixKP)
}
func TestRouteWithoutAnOriginASGivesOnlyThePrefixTag(t *testing.T) {
t.Parallel()
url := startRouteWatch(t, map[string]standInRoute{fromDE: {prefixDE, 0}})
s, _, server := startWithClock(t, "", map[string]string{
lookupSource: off, routeWatchURL: url,
})
waitForRouteWatch(t, server, fromDE, 32)
wantTags(t, s.get(fromDE, http.StatusOK, requestlog.ActionForward),
"prefix="+prefixDE)
}
func TestUnsetRouteWatchChangesNothing(t *testing.T) {
t.Parallel()
s, _, server := startWithClock(t, "", map[string]string{lookupSource: off})
if server.RouteWatch != nil {
t.Error("RouteWatch is set while SWW_ROUTEWATCH_URL is unset")
}
line := s.get(fromDE, http.StatusOK, requestlog.ActionForward)
if _, present := line.fields[tagsMissing]; present {
t.Errorf("log line has %s", tagsMissing)
}
if _, present := line.fields["tags"]; present {
t.Errorf("log line has tags %v", line.fields["tags"])
}
}
func TestUnreachableRouteWatchLetsTheRequestThroughAndSaysWhy(t *testing.T) {
t.Parallel()
// A stand-in that is closed at once, so that the port refuses connections.
dead := httptest.NewServer(http.NotFoundHandler())
url := dead.URL
dead.Close()
s, _, _ := startWithClock(t, "", map[string]string{
lookupSource: off, routeWatchURL: url,
})
wantTagsMissing(t, s.get(fromDE, http.StatusOK, requestlog.ActionForward),
"no_answer_yet")
deadline := time.Now().Add(5 * time.Second)
for {
line := s.get(fromDE, http.StatusOK, requestlog.ActionForward)
missing, _ := line.fields[tagsMissing].(string)
if missing == "request_failed" {
return
}
if time.Now().After(deadline) {
t.Fatalf("still %q, want request_failed", missing)
}
time.Sleep(time.Millisecond)
}
}
func TestAddressWithoutARouteHasNoTagsAndTheLineSaysSo(t *testing.T) {
t.Parallel()
s, _, _ := startWithClock(t, "", map[string]string{
lookupSource: off, routeWatchURL: startRouteWatch(t, nil),
})
deadline := time.Now().Add(5 * time.Second)
for {
line := s.get(fromDE, http.StatusOK, requestlog.ActionForward)
missing, _ := line.fields[tagsMissing].(string)
if missing == "no_route" {
return
}
if time.Now().After(deadline) {
t.Fatalf("still %q, want no_route", missing)
}
time.Sleep(time.Millisecond)
}
}
func TestTagDenyRefusesOnARouteWatchPrefix(t *testing.T) {
t.Parallel()
url := startRouteWatch(t, map[string]standInRoute{
fromDE: {prefixDE, 64496}, fromKP: {prefixKP, 64511},
})
s, _, server := startWithClock(t, "", map[string]string{
lookupSource: off, routeWatchURL: url, tagDeny: "prefix=" + prefixKP,
})
waitForRouteWatch(t, server, fromKP, 32)
waitForRouteWatch(t, server, fromDE, 32)
wantHeuristic(t, s.get(fromKP, http.StatusForbidden, requestlog.ActionClientDenied),
"tag_deny", `{"tag": "prefix=`+prefixKP+`"}`)
s.get(fromDE, http.StatusOK, requestlog.ActionForward)
}
func TestTagLimitPercentWorksWithAnIPv6Prefix(t *testing.T) {
t.Parallel()
url := startRouteWatch(t, map[string]standInRoute{fromV6: {prefixV6, 64496}})
s, _, server := startWithClock(t, "", map[string]string{
lookupSource: off, routeWatchURL: url, rateLimitPerMinute: "2",
tagLimitPercent: "prefix=" + prefixV6 + ":50",
})
waitForRouteWatch(t, server, fromV6, 64)
// Half of two is one: the second request breaks the limit.
line := s.get(fromV6, http.StatusOK, requestlog.ActionForward)
wantTags(t, line, "asn=64496", "prefix="+prefixV6)
s.get(fromV6, http.StatusForbidden, requestlog.ActionClientIsRateLimited)
}
-4
View File
@@ -187,10 +187,6 @@ type Line struct {
Reputation []string `json:"reputation,omitempty"`
// Tags are the request's tags, sorted.
Tags []string `json:"tags,omitempty"`
// RouteWatchTagsMissing, while SWW_ROUTEWATCH_URL is set, says why the
// request got no tags from routewatch: no_answer_yet, request_failed or
// no_route.
RouteWatchTagsMissing string `json:"routewatch_tags_missing,omitempty"`
// Offence is the offence the request was held as, OffenceLimit.
Offence string `json:"offence,omitempty"`
// BanExpires is when the ban the request made, or was refused under,
-243
View File
@@ -1,243 +0,0 @@
// Package routewatch asks a routewatch instance about client addresses, in the
// background, and keeps its answers in memory. A routewatch instance knows, for
// an address, the netblock it is announced in and the AS that originates it.
package routewatch
import (
"context"
"encoding/json"
"errors"
"fmt"
"io"
"log/slog"
"net/http"
"net/netip"
"net/url"
"sync"
"time"
"github.com/hashicorp/golang-lru/v2/simplelru"
)
const (
// answerKept is how long an answer is used after it came.
answerKept = time.Hour
// failureKept is how long routewatch is not asked again after a query
// fails, so that a dead routewatch is not asked on every request.
failureKept = time.Minute
// queryTimeout is how long one query may take. routewatch answers only
// after a reverse DNS lookup it allows 2 seconds, so this is longer.
queryTimeout = 5 * time.Second
// maxAnswers is how many answers are kept. Past it, the one that came
// longest ago is dropped.
maxAnswers = 100000
// maxQueries is how many queries may be under way at once. Past it, a
// client is not asked about until its next request, so that a swarm of
// new addresses cannot fill the memory.
maxQueries = 100
// maxAnswerBytes is the most of an answer that is read.
maxAnswerBytes = 64 << 10
)
// Missing is why Route gives no route. The request log's field
// routewatch_tags_missing gives these values.
type Missing string
const (
// NoAnswerYet is a client routewatch has not answered about yet.
NoAnswerYet Missing = "no_answer_yet"
// RequestFailed is no answer because a recent query failed.
RequestFailed Missing = "request_failed"
// NoRoute is routewatch knowing no live route for the address.
NoRoute Missing = "no_route"
)
var (
errStatus = errors.New("routewatch answered")
errNoResult = errors.New("the answer holds no result")
errNotNetblock = errors.New("the answer's netblock is not a netblock")
)
// Route is what routewatch knows of an address: the netblock it is announced
// in, and the AS that originates it.
type Route struct {
Netblock netip.Prefix
ASN uint32
}
// Params are what New needs.
type Params struct {
// URL is where routewatch is (SWW_ROUTEWATCH_URL).
URL string
// Now tells the time, normally time.Now in UTC.
Now func() time.Time
// ProcessLog receives the query that fails, and why.
ProcessLog *slog.Logger
}
// Client asks a routewatch instance about clients, in the background, and keeps
// its answers. It is safe for concurrent use.
type Client struct {
params Params
httpClient *http.Client
mu sync.Mutex
// answers are by client. Each is added as it comes and never moved up,
// so that the one that came longest ago is the first dropped.
answers *simplelru.LRU[netip.Prefix, answer]
// asking are the clients whose query is under way.
asking map[netip.Prefix]bool
// retryAt is when routewatch may be asked again after a query failed.
retryAt time.Time
}
// answer is one kept answer: the route, or why there is none, and when it came.
type answer struct {
route Route
missing Missing
came time.Time
}
// New returns a Client with no answer yet.
func New(params Params) *Client {
answers, err := simplelru.NewLRU[netip.Prefix, answer](maxAnswers, nil)
if err != nil {
panic(err) // NewLRU fails only for a size below one
}
return &Client{
params: params,
httpClient: &http.Client{},
answers: answers,
asking: map[netip.Prefix]bool{},
}
}
// Route returns what routewatch answered about client, an IPv4 address or an
// IPv6 group, or why it has no route. It never waits for a query. A client
// without an answer is asked about in the background, by addr, the address its
// request came from, unless its query is under way, a query failed less than
// failureKept ago, or maxQueries are under way. ctx is the context of the
// client's request, and a query goes on after the request ends.
func (c *Client) Route(
ctx context.Context, client netip.Prefix, addr netip.Addr,
) (Route, Missing) {
c.mu.Lock()
defer c.mu.Unlock()
now := c.params.Now()
kept, found := c.answers.Peek(client)
if found && now.Sub(kept.came) < answerKept {
return kept.route, kept.missing
}
if now.Before(c.retryAt) {
return Route{}, RequestFailed
}
if !c.asking[client] && len(c.asking) < maxQueries {
c.asking[client] = true
go c.ask(context.WithoutCancel(ctx), client, addr)
}
return Route{}, NoAnswerYet
}
// ask asks routewatch about addr, keeps the answer as client's, and notes the
// query as no longer under way. A query that fails gives no answer, and
// routewatch is not asked again for failureKept. Only the first failure of
// those under way is logged.
func (c *Client) ask(ctx context.Context, client netip.Prefix, addr netip.Addr) {
route, missing, err := c.query(ctx, addr)
now := c.params.Now()
c.mu.Lock()
delete(c.asking, client)
firstFailure := err != nil && !now.Before(c.retryAt)
if err == nil {
c.answers.Add(client, answer{route: route, missing: missing, came: now})
} else {
c.retryAt = now.Add(failureKept)
}
c.mu.Unlock()
if firstFailure {
c.params.ProcessLog.Warn("asking routewatch failed", "error", err.Error())
}
}
// query asks routewatch's IP info API about addr. An answer other than 200, one
// that cannot be read, and none within queryTimeout, fail. A result with no
// netblock is an address without a live route.
func (c *Client) query(ctx context.Context, addr netip.Addr) (Route, Missing, error) {
ctx, cancel := context.WithTimeout(ctx, queryTimeout)
defer cancel()
endpoint, err := url.JoinPath(c.params.URL, "api/v1/ip", addr.String())
if err != nil {
return Route{}, "", fmt.Errorf("make the URL: %w", err)
}
req, err := http.NewRequestWithContext(ctx, http.MethodGet, endpoint, http.NoBody)
if err != nil {
return Route{}, "", fmt.Errorf("make the request: %w", err)
}
req.Header.Set("Accept", "application/json")
res, err := c.httpClient.Do(req)
if err != nil {
// Do's error names the URL, which holds the client's address, which
// is not to be logged: only what went wrong is kept.
return Route{}, "", fmt.Errorf("ask routewatch: %w", errors.Unwrap(err))
}
defer func() {
_ = res.Body.Close()
}()
if res.StatusCode != http.StatusOK {
return Route{}, "", fmt.Errorf("%w %s", errStatus, res.Status)
}
return readAnswer(io.LimitReader(res.Body, maxAnswerBytes))
}
// readAnswer reads routewatch's answer about one address.
func readAnswer(body io.Reader) (Route, Missing, error) {
var decoded struct {
Data struct {
Results []struct {
Netblock string `json:"netblock"`
ASN uint32 `json:"asn"`
} `json:"results"`
} `json:"data"`
}
err := json.NewDecoder(body).Decode(&decoded)
if err != nil {
return Route{}, "", fmt.Errorf("read the answer: %w", err)
}
if len(decoded.Data.Results) == 0 {
return Route{}, "", errNoResult
}
result := decoded.Data.Results[0]
if result.Netblock == "" {
return Route{}, NoRoute, nil
}
netblock, err := netip.ParsePrefix(result.Netblock)
if err != nil {
return Route{}, "", fmt.Errorf("%w: %q", errNotNetblock, result.Netblock)
}
return Route{Netblock: netblock.Masked(), ASN: result.ASN}, "", nil
}
-333
View File
@@ -1,333 +0,0 @@
package routewatch_test
import (
"bytes"
"context"
"log/slog"
"net/http"
"net/http/httptest"
"net/netip"
"strings"
"sync"
"sync/atomic"
"testing"
"time"
"sneak.berlin/go/smallwebwaf/internal/routewatch"
)
const (
clientAddr = "192.0.2.1"
routedAnswer = `{"status":"ok","data":{"results":[` +
`{"ip":"192.0.2.1","netblock":"192.0.2.0/24","asn":64496}]}}`
noRouteAnswer = `{"status":"ok","data":{"results":[{"ip":"192.0.2.1"}]}}`
// slowAnswer is longer than routewatch allows a reverse DNS lookup.
slowAnswer = 2500 * time.Millisecond
// waitLimit is longer than the Client allows a query.
waitLimit = 10 * time.Second
)
// clock is a time the tests move by hand.
type clock struct {
mu sync.Mutex
now time.Time
}
func (c *clock) Now() time.Time {
c.mu.Lock()
defer c.mu.Unlock()
return c.now
}
func (c *clock) advance(by time.Duration) {
c.mu.Lock()
defer c.mu.Unlock()
c.now = c.now.Add(by)
}
// fixture is a Client asking a stand-in for routewatch. queries counts what the
// stand-in was asked, paths holds the last path it was asked for, and log holds
// the process log.
type fixture struct {
t *testing.T
server *httptest.Server
client *routewatch.Client
clk *clock
queries *atomic.Int32
path *atomic.Value
log *syncBuffer
}
// syncBuffer is a buffer the background queries can log to while a test reads it.
type syncBuffer struct {
mu sync.Mutex
buf bytes.Buffer
}
func (b *syncBuffer) Write(p []byte) (int, error) {
b.mu.Lock()
defer b.mu.Unlock()
return b.buf.Write(p)
}
func (b *syncBuffer) String() string {
b.mu.Lock()
defer b.mu.Unlock()
return b.buf.String()
}
func newFixture(t *testing.T, standIn http.HandlerFunc) *fixture {
t.Helper()
f := &fixture{
t: t,
clk: &clock{now: time.Date(2026, 11, 9, 0, 0, 0, 0, time.UTC)},
queries: &atomic.Int32{},
path: &atomic.Value{},
log: &syncBuffer{},
}
f.server = httptest.NewServer(http.HandlerFunc(
func(w http.ResponseWriter, r *http.Request) {
f.queries.Add(1)
f.path.Store(r.URL.Path)
standIn(w, r)
}))
t.Cleanup(f.server.Close)
f.client = routewatch.New(routewatch.Params{
URL: f.server.URL + "/", Now: f.clk.Now,
ProcessLog: slog.New(slog.NewTextHandler(f.log, nil)),
})
return f
}
func answering(body string) http.HandlerFunc {
return func(w http.ResponseWriter, _ *http.Request) {
_, _ = w.Write([]byte(body))
}
}
func (f *fixture) route() (routewatch.Route, routewatch.Missing) {
addr := netip.MustParseAddr(clientAddr)
client := netip.PrefixFrom(addr, addr.BitLen())
return f.client.Route(context.Background(), client, addr)
}
// waitFor asks until the answer is why, or, while why is "", until there is a
// route, and returns the route.
func (f *fixture) waitFor(why routewatch.Missing) routewatch.Route {
f.t.Helper()
deadline := time.Now().Add(waitLimit)
for {
route, missing := f.route()
if missing == why {
return route
}
if time.Now().After(deadline) {
f.t.Fatalf("still %q after %s, want %q", missing, waitLimit, why)
}
time.Sleep(time.Millisecond)
}
}
// wantFailureLoggedWithoutTheClient waits for the process log to name a failed
// query, and checks that it does not name the client.
func (f *fixture) wantFailureLoggedWithoutTheClient() {
f.t.Helper()
// The failure is noted before it is logged.
deadline := time.Now().Add(waitLimit)
for !strings.Contains(f.log.String(), "asking routewatch failed") {
if time.Now().After(deadline) {
f.t.Fatalf("process log %q names no failure", f.log.String())
}
time.Sleep(time.Millisecond)
}
if strings.Contains(f.log.String(), clientAddr) {
f.t.Errorf("process log %q names the client", f.log.String())
}
}
func TestRouteGivesRoutewatchsAnswerAndAsksByAddress(t *testing.T) {
t.Parallel()
f := newFixture(t, answering(routedAnswer))
if _, missing := f.route(); missing != routewatch.NoAnswerYet {
t.Errorf("before the answer: %q, want %q", missing, routewatch.NoAnswerYet)
}
route := f.waitFor("")
want := routewatch.Route{Netblock: netip.MustParsePrefix("192.0.2.0/24"), ASN: 64496}
if route != want {
t.Errorf("route %+v, want %+v", route, want)
}
if got := f.path.Load(); got != "/api/v1/ip/"+clientAddr {
t.Errorf("asked for path %v", got)
}
}
func TestAnswerAfterASlowReverseDNSLookupIsKept(t *testing.T) {
t.Parallel()
f := newFixture(t, func(w http.ResponseWriter, _ *http.Request) {
time.Sleep(slowAnswer)
_, _ = w.Write([]byte(routedAnswer))
})
f.route()
route := f.waitFor("")
want := routewatch.Route{Netblock: netip.MustParsePrefix("192.0.2.0/24"), ASN: 64496}
if route != want {
t.Errorf("route %+v, want %+v", route, want)
}
}
func TestAddressWithoutARouteIsKeptAsNoRoute(t *testing.T) {
t.Parallel()
f := newFixture(t, answering(noRouteAnswer))
f.route()
f.waitFor(routewatch.NoRoute)
if got := f.queries.Load(); got != 1 {
t.Errorf("%d queries, want 1", got)
}
}
func TestAnswerRoutewatchCannotBeTakenFromFails(t *testing.T) {
t.Parallel()
tests := map[string]http.HandlerFunc{
"500": func(w http.ResponseWriter, _ *http.Request) {
w.WriteHeader(http.StatusInternalServerError)
},
"not JSON": answering("no"),
"no results": answering(`{"status":"ok","data":{"results":[]}}`),
"bad netblock": answering(
`{"status":"ok","data":{"results":[{"netblock":"nowhere","asn":1}]}}`),
}
for name, standIn := range tests {
t.Run(name, func(t *testing.T) {
t.Parallel()
f := newFixture(t, standIn)
f.route()
f.waitFor(routewatch.RequestFailed)
f.wantFailureLoggedWithoutTheClient()
})
}
}
func TestRefusedConnectionFailsTheQueryAndTheLogOmitsTheClient(t *testing.T) {
t.Parallel()
f := newFixture(t, answering(routedAnswer))
f.server.Close()
f.route()
f.waitFor(routewatch.RequestFailed)
f.wantFailureLoggedWithoutTheClient()
}
func TestFailureIsKeptForAMinuteSoRoutewatchIsNotAskedPerRequest(t *testing.T) {
t.Parallel()
f := newFixture(t, func(w http.ResponseWriter, _ *http.Request) {
w.WriteHeader(http.StatusInternalServerError)
})
f.route()
f.waitFor(routewatch.RequestFailed)
f.route()
if got := f.queries.Load(); got != 1 {
t.Errorf("%d queries while the failure is kept, want 1", got)
}
f.clk.advance(time.Minute)
f.route()
deadline := time.Now().Add(waitLimit)
for f.queries.Load() != 2 {
if time.Now().After(deadline) {
t.Fatalf("%d queries after the failure expired, want 2", f.queries.Load())
}
time.Sleep(time.Millisecond)
}
}
func TestHangingRoutewatchNeverDelaysAndFailsTheQuery(t *testing.T) {
t.Parallel()
f := newFixture(t, func(_ http.ResponseWriter, r *http.Request) {
<-r.Context().Done()
})
start := time.Now()
f.route()
if elapsed := time.Since(start); elapsed > time.Second {
t.Errorf("Route took %s while routewatch hangs", elapsed)
}
f.waitFor(routewatch.RequestFailed)
f.wantFailureLoggedWithoutTheClient()
}
func TestRepeatedRequestsMakeOneQueryAndTheAnswerExpires(t *testing.T) {
t.Parallel()
release := make(chan struct{})
f := newFixture(t, func(w http.ResponseWriter, _ *http.Request) {
<-release
_, _ = w.Write([]byte(routedAnswer))
})
for range 50 {
f.route()
}
close(release)
f.waitFor("")
for range 50 {
f.route()
}
if got := f.queries.Load(); got != 1 {
t.Errorf("%d queries for repeated requests, want 1", got)
}
f.clk.advance(time.Hour)
f.waitFor(routewatch.NoAnswerYet)
f.waitFor("")
if got := f.queries.Load(); got != 2 {
t.Errorf("%d queries after the answer expired, want 2", got)
}
}