Compare commits

..
Author SHA1 Message Date
sneak 2284398c81 rproxy tests: leave room for a busy host in the 408 test
check / check (push) Waiting to run
TestAVisitorTooSlowToSendTheRestIsAnswered408 still used a 3-second
request time limit. Under load, rproxy had not yet started waiting for
the rest of the body when it ran out, and answered 504 instead of 408,
as reported on the issue. It now uses 5 seconds, like the other
time-limit tests in this PR.

Model: opus-5-5
2026-11-19 18:45:49 +00:00
sneak 7002a5bfb0 rproxy tests: leave room for a busy host in five timing tests (closes #451)
check / check (push) Waiting to run
Three tests that wait out a 2-second request or response time limit now
use 5 seconds, so the request reaches the app before it runs out under
load. The WebSocket one also renews its connection's deadline after the
wait.

The two breaker tests keep sending requests until one is kept from
asking waf, instead of expecting the sixth to be. On a busy host the five
asks can spread over more than five seconds, or the sixth can arrive
after the open breaker's one-second wait. Their stand-ins for waf now
fail every ask, so a request that asks never closes the breaker.

Model: opus-5-5
2026-11-19 18:34:21 +00:00
clawbot 11ac8aee50 config: export the setting names waf logs (closes #448)
check / check (push) Waiting to run
waf wrote eight setting names as string literals into log lines and
ban notes, and config wrote each again where it reads the setting. A
rename that missed one would log a setting that does not exist. config
now exports a constant for each name, used by its reader and by waf.

Model: opus-5-5
2026-11-19 19:29:49 +01:00
clawbot e7cc0adbdb waf: end the asks to GeoJS, the DNSBL zones, AbuseIPDB and routewatch at the stop (closes #439)
check / check (push) Waiting to run
Each of the four keeps the cancel function of every ask it started, and a
wait group of their goroutines. Its new method (GeoJS.StopAsking,
DNSBL.StopAsking, AbuseIPDB.StopChecking, routewatch Client.StopAsking)
refuses further asks, ends those under way, and returns once their
goroutines have. An ask ended that way is no failure: no alert, no failure
count, no wait before the next. Each provider in internal/sww adds the
method as a stop hook, so the last write of the state files follows it.

Model: opus-5-5
2026-11-19 19:02:12 +01:00
clawbot 805c64a256 config: give the fixed-value settings named types (closes #427)
check / check (push) Waiting to run
BytesCounted, LookupSource, BlocklistAction, DNSBLAbuseIPDBAction and
CoreRuleSetMode become named string types declared in config, with
constants for their values. Every comparison in waf and sww uses the
constants, and the private duplicates fileSource and deny go. The two
action settings share ReputationAction. The values and the strings
written to logs, metrics and alerts are unchanged.

Model: opus-5-5
2026-11-19 18:32:29 +01:00
28 changed files with 923 additions and 815 deletions

No files matched your search

+122 -48
View File
@@ -27,6 +27,19 @@ import (
"sneak.berlin/go/smallwebwaf/internal/remotelog"
)
// The names of the settings that waf writes into log lines and ban notes,
// as the setting that refused a request or lowered a client's limits.
const (
ExclusivelyAllowedCountriesSetting = "SWW_WAF_EXCLUSIVELY_ALLOWED_COUNTRIES"
TagLimitPercentSetting = "SWW_WAF_TAG_LIMIT_PERCENT"
ASNBytesLimitPercentSetting = "SWW_WAF_ASN_BYTES_LIMIT_PERCENT"
CountryBytesLimitPercentSetting = "SWW_WAF_COUNTRY_BYTES_LIMIT_PERCENT"
UnknownCountryLimitPercentSetting = "SWW_WAF_UNKNOWN_COUNTRY_LIMIT_PERCENT"
ASNLimitPercentURLSetting = "SWW_WAF_ASN_LIMIT_PERCENT_URL"
BlocklistActionSetting = "SWW_WAF_BLOCKLIST_ACTION"
DNSBLAbuseIPDBActionSetting = "SWW_WAF_DNSBL_ABUSEIPDB_ACTION"
)
// WAF is the settings of `sww waf`. A time or size limit of zero is off.
type WAF struct {
// ListenAddr is where waf listens (SWW_WAF_LISTEN_ADDR).
@@ -75,7 +88,7 @@ type WAF struct {
BytesLimitPerMinute int64
BytesLimitPerHour int64
BytesLimitPerDay int64
BytesCounted string
BytesCounted BytesCounted
// LookupSource is where each client's AS number and country are looked
// up (SWW_WAF_LOOKUP_SOURCE): geojs, file, or off for nowhere.
// LookupDBPath is the lookup database, the IPinfo Lite file looked up in
@@ -86,7 +99,7 @@ type WAF struct {
// AddLookupHeaders is true when the app is passed the client's AS number
// and country in headers (SWW_WAF_ADD_LOOKUP_HEADERS), by default unless
// LookupSource is off.
LookupSource string
LookupSource LookupSource
LookupDBPath string
ASNAndCountryLookupTimeout time.Duration
AddLookupHeaders bool
@@ -131,7 +144,7 @@ type WAF struct {
// BlocklistLimitPercent is the percentage of every limit it gets.
BlocklistURLs []string
BlocklistRefresh time.Duration
BlocklistAction string
BlocklistAction ReputationAction
BlocklistLimitPercent int64
// DNSBLZones are the DNSBL zones clients are asked about
// (SWW_WAF_DNSBL_ZONES), through DNSBLResolver (SWW_WAF_DNSBL_RESOLVER), or
@@ -154,7 +167,7 @@ type WAF struct {
AbuseIPDBKey string
AbuseIPDBMinScore int64
AbuseIPDBDailyBudget int
DNSBLAbuseIPDBAction string
DNSBLAbuseIPDBAction ReputationAction
DNSBLAbuseIPDBLimitPercent int64
IPReputationCacheTTL time.Duration
DNSBLAbuseIPDBTimeout time.Duration
@@ -237,7 +250,7 @@ type WAF struct {
// reads none.
// CoreRuleSetExclusionsFile is the path of the file of its exclusions
// (SWW_WAF_CORE_RULE_SET_EXCLUSIONS_FILE), "" while there is none.
CoreRuleSetMode string
CoreRuleSetMode CoreRuleSetMode
CoreRuleSetParanoiaLevel int
CoreRuleSetScoreThreshold int
CoreRuleSetDisabledRules []int
@@ -316,20 +329,66 @@ type WAF struct {
settings []slog.Attr
}
// BytesCounted is which body bytes count toward the byte limits and the
// byte anomaly thresholds, a value of SWW_WAF_BYTES_COUNTED.
type BytesCounted string
// The values of SWW_WAF_BYTES_COUNTED.
const (
// BytesCountedResponse counts the response's body bytes.
BytesCountedResponse BytesCounted = "response"
// BytesCountedRequest counts the request's body bytes.
BytesCountedRequest BytesCounted = "request"
// BytesCountedBoth counts the body bytes of both.
BytesCountedBoth BytesCounted = "both"
)
// LookupSource is where each client's AS number and country are looked up,
// a value of SWW_WAF_LOOKUP_SOURCE.
type LookupSource string
// The values of SWW_WAF_LOOKUP_SOURCE.
const (
// LookupSourceGeoJS asks GeoJS.
LookupSourceGeoJS LookupSource = "geojs"
// LookupSourceFile looks clients up in the lookup database, the file
// SWW_WAF_LOOKUP_DB_PATH names.
LookupSourceFile LookupSource = "file"
// LookupSourceOff looks no client up.
LookupSourceOff LookupSource = "off"
)
// ReputationAction is what is done with a client a reputation source
// lists, a value of SWW_WAF_BLOCKLIST_ACTION and of
// SWW_WAF_DNSBL_ABUSEIPDB_ACTION.
type ReputationAction string
// The values of SWW_WAF_BLOCKLIST_ACTION and SWW_WAF_DNSBL_ABUSEIPDB_ACTION.
const (
// ReputationActionDeny refuses the client's requests.
ReputationActionDeny ReputationAction = "deny"
// ReputationActionLimit gives the client a percentage of every limit,
// written limit:<percent> in the setting.
ReputationActionLimit ReputationAction = "limit"
// ReputationActionLog only notes the hit in the log line and raises an
// alert.
ReputationActionLog ReputationAction = "log"
)
// CoreRuleSetMode is what the Core Rule Set does, a value of
// SWW_WAF_CORE_RULE_SET_MODE.
type CoreRuleSetMode string
// The values of SWW_WAF_CORE_RULE_SET_MODE.
const (
// CoreRuleSetModeOff runs no request through the Core Rule Set.
CoreRuleSetModeOff = off
CoreRuleSetModeOff CoreRuleSetMode = "off"
// CoreRuleSetModeDetect logs and alerts a match, and refuses nothing.
CoreRuleSetModeDetect = "detect"
CoreRuleSetModeDetect CoreRuleSetMode = "detect"
// CoreRuleSetModeBlock refuses a match with 403.
CoreRuleSetModeBlock = "block"
CoreRuleSetModeBlock CoreRuleSetMode = "block"
)
// fileSource is the SWW_WAF_LOOKUP_SOURCE that looks clients up in the
// lookup database, the file SWW_WAF_LOOKUP_DB_PATH names.
const fileSource = "file"
const (
day = 24 * time.Hour
ipv4Bits = 32
@@ -558,10 +617,13 @@ func WAFFromEnvironment(v *viper.Viper) (*WAF, error) {
),
BytesLimitPerHour: env.size("SWW_WAF_BYTES_LIMIT_PER_HOUR", "50G"),
BytesLimitPerDay: env.size("SWW_WAF_BYTES_LIMIT_PER_DAY", "100G"),
BytesCounted: env.bytesCounted("SWW_WAF_BYTES_COUNTED", "both"),
BytesCounted: env.bytesCounted(
"SWW_WAF_BYTES_COUNTED",
BytesCountedBoth,
),
LookupSource: env.lookupSource(
"SWW_WAF_LOOKUP_SOURCE",
"geojs",
LookupSourceGeoJS,
),
LookupDBPath: env.value("SWW_WAF_LOOKUP_DB_PATH", ""),
ASNAndCountryLookupTimeout: env.milliseconds(
@@ -569,24 +631,24 @@ func WAFFromEnvironment(v *viper.Viper) (*WAF, error) {
"800",
),
ExclusivelyAllowedCountries: env.countries(
"SWW_WAF_EXCLUSIVELY_ALLOWED_COUNTRIES", ""),
ExclusivelyAllowedCountriesSetting, ""),
TagLists: env.tagLists("SWW_WAF_TAG_LISTS"),
TagDeny: env.tags("SWW_WAF_TAG_DENY"),
TagLimitPercent: env.tagPercents("SWW_WAF_TAG_LIMIT_PERCENT"),
TagLimitPercent: env.tagPercents(TagLimitPercentSetting),
RouteWatchURL: env.routeWatchURL("SWW_WAF_ROUTEWATCH_URL"),
ASNBytesLimitPercent: env.percents(
"SWW_WAF_ASN_BYTES_LIMIT_PERCENT",
ASNBytesLimitPercentSetting,
ParseASN,
),
CountryBytesLimitPercent: env.percents(
"SWW_WAF_COUNTRY_BYTES_LIMIT_PERCENT",
CountryBytesLimitPercentSetting,
parseCountry,
),
UnknownCountryLimitPercent: env.percent(
"SWW_WAF_UNKNOWN_COUNTRY_LIMIT_PERCENT",
UnknownCountryLimitPercentSetting,
"100",
),
ASNLimitPercentURL: env.listURL("SWW_WAF_ASN_LIMIT_PERCENT_URL"),
ASNLimitPercentURL: env.listURL(ASNLimitPercentURLSetting),
BlocklistURLs: env.listURLs("SWW_WAF_BLOCKLIST_URLS"),
BlocklistRefresh: env.refresh("SWW_WAF_BLOCKLIST_REFRESH", "24h"),
DNSBLZones: env.zones("SWW_WAF_DNSBL_ZONES"),
@@ -758,11 +820,11 @@ func WAFFromEnvironment(v *viper.Viper) (*WAF, error) {
cfg.InstanceName, cfg.LogRemoteURL != nil)
// The app gets the headers by default, unless no client is looked up.
cfg.AddLookupHeaders = env.boolean("SWW_WAF_ADD_LOOKUP_HEADERS",
strconv.FormatBool(cfg.LookupSource != off))
strconv.FormatBool(cfg.LookupSource != LookupSourceOff))
cfg.BlocklistAction, cfg.BlocklistLimitPercent = env.action(
"SWW_WAF_BLOCKLIST_ACTION", "deny")
BlocklistActionSetting, string(ReputationActionDeny))
cfg.DNSBLAbuseIPDBAction, cfg.DNSBLAbuseIPDBLimitPercent = env.action(
"SWW_WAF_DNSBL_ABUSEIPDB_ACTION", "limit:25")
DNSBLAbuseIPDBActionSetting, "limit:25")
env.checkInstanceNameForNtfy(cfg.InstanceName, cfg.AlertNtfyURL != nil)
env.checkLookupDBPath(cfg)
@@ -858,9 +920,12 @@ func (e *environment) count(name, defaultValue string) int64 {
// bytesCounted reads the setting that is which body bytes count toward the
// byte limits and the byte anomaly thresholds: response, request or both.
func (e *environment) bytesCounted(name, defaultValue string) string {
value := e.value(name, defaultValue)
if value != "response" && value != "request" && value != "both" {
func (e *environment) bytesCounted(
name string, defaultValue BytesCounted,
) BytesCounted {
value := BytesCounted(e.value(name, string(defaultValue)))
if value != BytesCountedResponse && value != BytesCountedRequest &&
value != BytesCountedBoth {
e.check(name, fmt.Errorf("%q %w", value, errNotBytesCounted))
}
@@ -905,8 +970,10 @@ func (e *environment) trapPaths(name string) []string {
// coreRuleSetMode reads the setting that is what the Core Rule Set does: off,
// detect or block.
func (e *environment) coreRuleSetMode(name, defaultValue string) string {
mode := e.value(name, defaultValue)
func (e *environment) coreRuleSetMode(
name string, defaultValue CoreRuleSetMode,
) CoreRuleSetMode {
mode := CoreRuleSetMode(e.value(name, string(defaultValue)))
if mode != CoreRuleSetModeOff && mode != CoreRuleSetModeDetect &&
mode != CoreRuleSetModeBlock {
e.check(name, fmt.Errorf("%q %w", mode, errNotCoreRuleSetMode))
@@ -1056,20 +1123,24 @@ func (e *environment) refresh(name, defaultValue string) time.Duration {
// 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" {
func (e *environment) action(
name, defaultValue string,
) (ReputationAction, int64) {
value := ReputationAction(e.value(name, defaultValue))
if value == ReputationActionDeny || value == ReputationActionLog {
return value, 0
}
percentText, isLimit := strings.CutPrefix(value, "limit:")
percentText, isLimit := strings.CutPrefix(
string(value), string(ReputationActionLimit)+":",
)
percent, err := ParsePercent(percentText)
if !isLimit || err != nil {
e.check(name, fmt.Errorf("%q %w", value, errNotAction))
}
return "limit", percent
return ReputationActionLimit, percent
}
// zones reads the setting that is the list of DNSBL zones. It is empty by
@@ -1120,9 +1191,12 @@ func (e *environment) routeWatchURL(name string) string {
// lookupSource reads the setting that is where clients are looked up:
// geojs, file, or off.
func (e *environment) lookupSource(name, defaultValue string) string {
source := e.value(name, defaultValue)
if source != "geojs" && source != fileSource && source != off {
func (e *environment) lookupSource(
name string, defaultValue LookupSource,
) LookupSource {
source := LookupSource(e.value(name, string(defaultValue)))
if source != LookupSourceGeoJS && source != LookupSourceFile &&
source != LookupSourceOff {
e.check(name, fmt.Errorf("%q %w", source, errNotLookupSource))
}
@@ -1134,12 +1208,12 @@ func (e *environment) lookupSource(name, defaultValue string) string {
// one source at a time.
func (e *environment) checkLookupDBPath(cfg *WAF) {
switch {
case cfg.LookupSource == fileSource && cfg.LookupDBPath == "":
case cfg.LookupSource == LookupSourceFile && cfg.LookupDBPath == "":
e.check("SWW_WAF_LOOKUP_SOURCE", fmt.Errorf(
"is file while SWW_WAF_LOOKUP_DB_PATH is unset; %w",
errNeedsDBPath,
))
case cfg.LookupSource != fileSource && cfg.LookupDBPath != "":
case cfg.LookupSource != LookupSourceFile && cfg.LookupDBPath != "":
e.check("SWW_WAF_LOOKUP_DB_PATH", fmt.Errorf(
"is set while SWW_WAF_LOOKUP_SOURCE is %s; %w",
cfg.LookupSource,
@@ -1159,12 +1233,12 @@ func (e *environment) checkLookupDBPath(cfg *WAF) {
func (e *environment) checkCountriesAndLookups(cfg *WAF) {
for _, country := range cfg.ExclusivelyAllowedCountries {
if slices.Contains(cfg.TagDeny, CountryTagPrefix+country) {
e.check("SWW_WAF_EXCLUSIVELY_ALLOWED_COUNTRIES",
e.check(ExclusivelyAllowedCountriesSetting,
fmt.Errorf("%q %w", country, errOnBothLists))
}
}
if cfg.LookupSource != off {
if cfg.LookupSource != LookupSourceOff {
return
}
@@ -1178,15 +1252,15 @@ func (e *environment) checkCountriesAndLookups(cfg *WAF) {
name string
set bool
}{
{"SWW_WAF_EXCLUSIVELY_ALLOWED_COUNTRIES", len(cfg.ExclusivelyAllowedCountries) > 0},
{ExclusivelyAllowedCountriesSetting, len(cfg.ExclusivelyAllowedCountries) > 0},
{"SWW_WAF_ADD_LOOKUP_HEADERS", cfg.AddLookupHeaders},
{"SWW_WAF_TAG_DENY", slices.ContainsFunc(cfg.TagDeny, onlyLookupGives)},
{"SWW_WAF_TAG_LIMIT_PERCENT", slices.ContainsFunc(
{TagLimitPercentSetting, slices.ContainsFunc(
slices.Collect(maps.Keys(cfg.TagLimitPercent)), onlyLookupGives)},
{"SWW_WAF_ASN_BYTES_LIMIT_PERCENT", len(cfg.ASNBytesLimitPercent) > 0},
{"SWW_WAF_COUNTRY_BYTES_LIMIT_PERCENT", len(cfg.CountryBytesLimitPercent) > 0},
{"SWW_WAF_UNKNOWN_COUNTRY_LIMIT_PERCENT", cfg.UnknownCountryLimitPercent < 100},
{"SWW_WAF_ASN_LIMIT_PERCENT_URL", cfg.ASNLimitPercentURL != ""},
{ASNBytesLimitPercentSetting, len(cfg.ASNBytesLimitPercent) > 0},
{CountryBytesLimitPercentSetting, len(cfg.CountryBytesLimitPercent) > 0},
{UnknownCountryLimitPercentSetting, cfg.UnknownCountryLimitPercent < 100},
{ASNLimitPercentURLSetting, cfg.ASNLimitPercentURL != ""},
{"SWW_WAF_ANOMALY_ASN_REQUESTS_PER_MINUTE", cfg.AnomalyASN.RequestsPerMinute > 0},
{"SWW_WAF_ANOMALY_ASN_REQUESTS_PER_HOUR", cfg.AnomalyASN.RequestsPerHour > 0},
{"SWW_WAF_ANOMALY_ASN_BYTES_PER_MINUTE", cfg.AnomalyASN.BytesPerMinute > 0},
@@ -1206,7 +1280,7 @@ func (e *environment) checkCountriesAndLookups(cfg *WAF) {
// blocklist too: the file at a URL is fetched as one list or the other.
func (e *environment) checkASNLimitPercentURL(cfg *WAF) {
if slices.Contains(cfg.BlocklistURLs, cfg.ASNLimitPercentURL) {
e.check("SWW_WAF_ASN_LIMIT_PERCENT_URL",
e.check(ASNLimitPercentURLSetting,
fmt.Errorf("%q %w", cfg.ASNLimitPercentURL, errInBlocklistURLs))
}
}
@@ -1233,7 +1307,7 @@ func (e *environment) checkTags(cfg *WAF) {
}
for tag := range cfg.TagLimitPercent {
e.check("SWW_WAF_TAG_LIMIT_PERCENT", checkTag(cfg, tag))
e.check(TagLimitPercentSetting, checkTag(cfg, tag))
}
}
+6 -5
View File
@@ -614,7 +614,7 @@ func TestCoreRuleSetSettings(t *testing.T) {
},
{
environment{
coreRuleSetMode: config.CoreRuleSetModeDetect,
coreRuleSetMode: string(config.CoreRuleSetModeDetect),
coreRuleSetParanoiaLevel: "4",
coreRuleSetScoreThreshold: "10",
coreRuleSetDisabledRules: "942100, 920350",
@@ -2033,8 +2033,9 @@ func TestReputationSettingsAsSet(t *testing.T) {
}
for _, tc := range []struct {
value, action string
percent int64
value string
action config.ReputationAction
percent int64
}{
{actionDeny, actionDeny, 0},
{actionLog, actionLog, 0},
@@ -2167,7 +2168,7 @@ func TestDNSBLSettingsAsSet(t *testing.T) {
// want are the zones, resolver, action and percent config.WAF gives.
want []string
wantResolver string
wantAction string
wantAction config.ReputationAction
wantPercent int64
}{
{
@@ -3049,7 +3050,7 @@ func TestLogsEachSettingWithItsValue(t *testing.T) {
logLevel: "info",
ruleFilesDir: "/etc/sww/rules.d",
ruleFilesEnabled: "true",
coreRuleSetMode: config.CoreRuleSetModeBlock,
coreRuleSetMode: string(config.CoreRuleSetModeBlock),
coreRuleSetParanoiaLevel: "1",
coreRuleSetScoreThreshold: "5",
coreRuleSetDisabledRules: defaultCoreRuleSetDisabledRules,
+47 -14
View File
@@ -99,8 +99,13 @@ type GeoJS struct {
// waiting are the clients without an answer: those to ask GeoJS about,
// and those it is being asked about.
waiting map[netip.Prefix]*wait
// asking is true while a request to GeoJS is under way.
asking bool
// asking is true while a request to GeoJS is under way, and endAsking
// ends it. stopped is true once StopAsking has been called, and
// askGoroutines waits for the goroutine that makes the requests.
asking bool
endAsking context.CancelFunc
stopped bool
askGoroutines sync.WaitGroup
// retryDelay is how long GeoJS is left alone after its last failure,
// zero after an answer; retryAt is when it may be asked again.
retryDelay time.Duration
@@ -210,6 +215,23 @@ func (g *GeoJS) LookUpWithoutWaiting(
return answer
}
// StopAsking ends the request to GeoJS under way, and returns once the
// goroutine that made it has. A request it ends is no failure: Failed is not
// called, and GeoJS is not left alone after it. After StopAsking, GeoJS is
// asked nothing.
func (g *GeoJS) StopAsking() {
g.mu.Lock()
g.stopped = true
if g.asking {
g.endAsking()
}
g.mu.Unlock()
g.askGoroutines.Wait()
}
// Kept returns client's answer, if one is kept, without asking GeoJS.
func (g *GeoJS) Kept(client netip.Prefix) (Answer, bool) {
g.mu.Lock()
@@ -294,7 +316,8 @@ func (g *GeoJS) answerOrWait(
}
if !g.asking {
// GeoJS is left alone after a failure, so no answer can come.
// GeoJS is left alone after a failure, or asked nothing after
// StopAsking, so no answer can come.
w.late = true
}
@@ -323,22 +346,28 @@ func (g *GeoJS) kept(client netip.Prefix) (Answer, bool) {
}
// ask starts asking GeoJS about the waiting clients, unless a request to
// it is under way or it is left alone after a failure. The requests to
// GeoJS are for every client waiting, so they go on when the caller whose
// ctx is given stops waiting.
// it is under way, it is left alone after a failure, or StopAsking has
// been called. The requests to GeoJS are for every client waiting, so they
// go on when the caller whose ctx is given stops waiting, until StopAsking.
func (g *GeoJS) ask(ctx context.Context) {
if g.asking || g.now().Before(g.retryAt) {
if g.asking || g.stopped || g.now().Before(g.retryAt) {
return
}
g.asking = true
askCtx, end := context.WithCancel(context.WithoutCancel(ctx))
g.asking, g.endAsking = true, end
go g.askAboutWaiting(context.WithoutCancel(ctx))
g.askGoroutines.Go(func() {
defer end()
g.askAboutWaiting(askCtx)
})
}
// askAboutWaiting asks GeoJS about the waiting clients, one request at a
// time, until none is left or GeoJS fails. Each answer kept is given to
// Answered, outside the lock, since Answered takes locks of its own.
// time, until none is left, GeoJS fails, or StopAsking ends ctx. Each
// answer kept is given to Answered, outside the lock, since Answered takes
// locks of its own.
func (g *GeoJS) askAboutWaiting(ctx context.Context) {
for {
clients := g.nextClients()
@@ -347,6 +376,10 @@ func (g *GeoJS) askAboutWaiting(ctx context.Context) {
}
given, err := g.request(ctx, clients)
if err != nil && ctx.Err() != nil {
continue // ended by StopAsking, which is no failure
}
kept, answered := g.keep(clients, given, err)
if g.answered != nil {
@@ -362,13 +395,13 @@ func (g *GeoJS) askAboutWaiting(ctx context.Context) {
}
// nextClients returns up to maxPerRequest of the waiting clients. When
// none is waiting, it returns none and notes that no request to GeoJS is
// under way.
// none is waiting, or StopAsking has been called, it returns none and
// notes that no request to GeoJS is under way.
func (g *GeoJS) nextClients() []netip.Prefix {
g.mu.Lock()
defer g.mu.Unlock()
if len(g.waiting) == 0 {
if len(g.waiting) == 0 || g.stopped {
g.asking = false
return nil
+38
View File
@@ -318,6 +318,44 @@ func TestCallerWhoseContextEndsEarlyDoesNotStopTheRequest(t *testing.T) {
})
}
func TestStopAskingEndsTheRequestUnderWayAsNoFailureAndAsksNoMore(
t *testing.T,
) {
t.Parallel()
synctest.Test(t, func(t *testing.T) {
service, _, g, reports := startWithReports()
clients := newClients()
service.set(hanging)
g.LookUpWithoutWaiting(t.Context(), clients())
waitForRequests(t, service, 1)
// The request is ended, not waited out, and is no failure.
began := time.Now()
g.StopAsking()
if waited := time.Since(began); waited != 0 {
t.Errorf("StopAsking returned after %s, want at once", waited)
}
reports.wantFailures(t)
// A new client is then not asked about, and its lookup does not
// wait.
service.set(answering)
wantCountry(t, g, clients(), "")
if waited := time.Since(began); waited != 0 {
t.Errorf("the lookup returned after %s, want at once", waited)
}
waitForRequests(t, service, 1)
reports.wantFailures(t)
})
}
func TestRequestOver10SecondsIsAFailure(t *testing.T) {
t.Parallel()
+266 -391
View File
@@ -16,33 +16,12 @@ import (
"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"
"sneak.berlin/go/smallwebwaf/internal/requestlog"
"sneak.berlin/go/smallwebwaf/internal/rules"
)
// Params are what New builds the metrics from. Each part is nil while it is
// off, and a part that is nil adds no metric of its own.
type Params struct {
// BusiestCount is how many countries and how many AS numbers get series
// of their own (SWW_WAF_METRICS_BUSIEST_COUNTRIES_AND_ASNS).
BusiestCount int
// InstanceName is the label instance every metric carries
// (SWW_WAF_INSTANCE_NAME).
InstanceName string
// Now returns the time at which Bans counts the bans active.
Now func() time.Time
Bans banLedger
Clients sized
Rules sized
Lists fetchedLists
DNSBL dnsblZones
AbuseIPDB abuseIPDBChecker
Alerts alertQueue
LookupFile lookupDatabase
RemoteLog remoteLogSender
}
// Metrics are sww's metrics. They are safe for concurrent use.
type Metrics struct {
// registry gives every metric registered with it the label instance.
@@ -58,10 +37,12 @@ type Metrics struct {
rateLimitHits *prometheus.CounterVec
offences *prometheus.CounterVec
coreRuleSetMatches *prometheus.CounterVec
ruleMatches *prometheus.CounterVec
reputationHits *prometheus.CounterVec
countries *busiest
asns *busiest
// ruleMatches are made by AddRules, and reputationHits by
// AddReputation.
ruleMatches *prometheus.CounterVec
reputationHits *prometheus.CounterVec
countries *busiest
asns *busiest
// GeoJSRequests are the requests to GeoJS, and GeoJSFailures those
// that failed. RequestsWithoutGeoJSAnswer are the requests that needed
@@ -79,18 +60,20 @@ type Metrics struct {
stateFileEditsSetAside *prometheus.CounterVec
}
// New returns the metrics, with the Go runtime's and the process's own, and
// those read from each part in params as the metrics are asked for.
// New returns the metrics, with the Go runtime's and the process's own.
// busiestCount is how many countries and how many AS numbers get series of
// their own (SWW_WAF_METRICS_BUSIEST_COUNTRIES_AND_ASNS). Every metric
// carries instanceName (SWW_WAF_INSTANCE_NAME) as its label instance.
//
//nolint:funlen // a few lines for each metric, a list that grows with them
func New(params Params) *Metrics {
func New(busiestCount int, instanceName string) *Metrics {
byStatus := []string{"status_class", "action"}
byFile := []string{"file"}
registry := prometheus.NewRegistry()
m := &Metrics{
registry: prometheus.WrapRegistererWith(
prometheus.Labels{"instance": params.InstanceName}, registry),
prometheus.Labels{"instance": instanceName}, registry),
handler: promhttp.HandlerFor(registry, promhttp.HandlerOpts{}),
inFlight: prometheus.NewGauge(prometheus.GaugeOpts{
Name: "sww_requests_in_flight",
@@ -131,19 +114,8 @@ func New(params Params) *Metrics {
"and the rule's id.",
[]string{"mode", "rule_id"},
),
ruleMatches: counterVec(
"sww_rule_matches_total",
"Requests that matched a rule of the rule files, by its id and action.",
[]string{"rule_id", "action"},
),
reputationHits: counterVec(
"sww_reputation_hits_total",
"Requests whose client a blocklist, a DNSBL zone or AbuseIPDB lists, "+
"by the list's URL, the zone, or abuseipdb.",
[]string{sourceLabel},
),
countries: newCountries(params.BusiestCount),
asns: newASNs(params.BusiestCount),
countries: newCountries(busiestCount),
asns: newASNs(busiestCount),
GeoJSRequests: prometheus.NewCounter(prometheus.CounterOpts{
Name: "sww_geojs_requests_total",
Help: "Requests to GeoJS.",
@@ -195,7 +167,6 @@ func New(params Params) *Metrics {
m.inFlight, m.requests, m.requestBytes, m.responseBytes,
m.requestDuration, m.upstreamDuration,
m.rateLimitHits, m.offences, m.coreRuleSetMatches,
m.ruleMatches, m.reputationHits,
m.countries, m.asns,
m.GeoJSRequests, m.GeoJSFailures, m.RequestsWithoutGeoJSAnswer,
m.stateFileWrites, m.stateFileWriteFailures,
@@ -203,45 +174,259 @@ func New(params Params) *Metrics {
m.stateFileEditsTakenIn, m.stateFileEditsSetAside,
)
if params.Bans != nil {
m.addBans(params.Bans, params.Now)
}
if params.Clients != nil {
m.addClients(params.Clients)
}
if params.Rules != nil {
m.addRules(params.Rules)
}
if params.Lists != nil {
m.addLists(params.Lists)
}
if params.DNSBL != nil {
m.addDNSBL(params.DNSBL)
}
if params.AbuseIPDB != nil {
m.addAbuseIPDB(params.AbuseIPDB)
}
if params.Alerts != nil {
m.addAlerts(params.Alerts)
}
if params.LookupFile != nil {
m.addLookupFile(params.LookupFile)
}
if params.RemoteLog != nil {
m.addRemoteLog(params.RemoteLog)
}
return m
}
// AddBansAndClients adds the metrics read from the ledger and the table
// of clients as the metrics are asked for: the bans made since the start,
// by cause, the bans active and permanent at now, and the clients in the
// table.
func (m *Metrics) AddBansAndClients(
ledger *bans.Ledger, limiter *ratelimit.Limiter, now func() time.Time,
) {
for _, cause := range []string{
bans.CauseLimit, bans.CauseAttack, bans.CauseAdmin, bans.CauseAbuseIPDB,
} {
m.registry.MustRegister(
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_bans_made_total",
Help: "Bans made, by cause.",
ConstLabels: prometheus.Labels{"cause": cause},
}, func() float64 {
return float64(ledger.Made(cause))
}),
)
}
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_active_bans",
Help: "Bans active now, the permanent ones included.",
}, func() float64 {
active, _ := ledger.Count(now())
return float64(active)
}),
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_permanent_bans",
Help: "Permanent bans not lifted.",
}, func() float64 {
_, permanent := ledger.Count(now())
return float64(permanent)
}),
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_tracked_clients",
Help: "Clients in the table of clients.",
}, func() float64 {
return float64(limiter.Len())
}),
)
}
// AddRules adds the metrics of the rule files: the requests that matched
// each rule, which RuleMatched counts, and the rules loaded from
// ruleFiles, read as the metrics are asked for. It is called once, before
// RuleMatched.
func (m *Metrics) AddRules(ruleFiles *rules.Files) {
m.ruleMatches = counterVec(
"sww_rule_matches_total",
"Requests that matched a rule of the rule files, by its id and action.",
[]string{"rule_id", "action"},
)
m.registry.MustRegister(m.ruleMatches,
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_rules_loaded",
Help: "Rules loaded from the rule files.",
}, func() float64 {
return float64(ruleFiles.Len())
}))
}
// AddRemoteLog adds the metrics of sending the log lines to
// SWW_WAF_LOG_REMOTE_URL, read from remote as the metrics are asked for: the
// lines sent, those dropped, and those waiting in the buffer.
func (m *Metrics) AddRemoteLog(remote *remotelog.Sender) {
m.registry.MustRegister(
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_remote_log_lines_sent_total",
Help: "Log lines sent to SWW_WAF_LOG_REMOTE_URL.",
}, func() float64 {
return float64(remote.Sent())
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_remote_log_lines_dropped_total",
Help: "Log lines dropped: the oldest in a full buffer, and those " +
"whose sending failed.",
}, func() float64 {
return float64(remote.Dropped())
}),
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_remote_log_lines_waiting",
Help: "Log lines in the buffer, waiting to be sent.",
}, func() float64 {
return float64(remote.Depth())
}),
)
}
// AddLookupFile adds the metrics of the lookup database, read as the
// metrics are asked for: when the file in use was read, which lastRead
// returns, and the replacements of it that could not be read, which
// readFailures returns.
func (m *Metrics) AddLookupFile(
lastRead func() time.Time,
readFailures func() int,
) {
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_lookup_database_last_read_timestamp_seconds",
Help: "When the lookup database in use was read, in seconds since 1970.",
}, func() float64 {
return float64(lastRead().Unix())
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_lookup_database_read_failures_total",
Help: "Replacements of the lookup database that could not be read.",
}, func() float64 {
return float64(readFailures())
}),
)
}
// sourceLabel is the label of the reputation metrics: a list's URL, a
// DNSBL zone, its key masked, or abuseipdb.
const sourceLabel = "source"
// 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, a
// zone's verdict or AbuseIPDB's score 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,
) {
m.reputationHits = counterVec(
"sww_reputation_hits_total",
"Requests whose client a blocklist, a DNSBL zone or AbuseIPDB lists, "+
"by the list's URL, the zone, or abuseipdb.",
[]string{sourceLabel},
)
m.registry.MustRegister(m.reputationHits)
for _, zone := range dnsbl.Zones() {
source := prometheus.Labels{sourceLabel: config.MaskZoneKey(zone)}
m.addReputationQueries(
source,
func() int { return dnsbl.Queries(zone) },
)
m.addReputationFailures(
source,
func() int { return dnsbl.Failures(zone) },
)
}
for _, listURL := range lists.URLs() {
source := prometheus.Labels{sourceLabel: listURL}
m.addReputationFailures(
source,
func() int { return lists.Failures(listURL) },
)
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_reputation_last_fetch_timestamp_seconds",
Help: "When the copy of the list in use was fetched, in seconds since " +
"1970, or 0 while there is none.",
ConstLabels: source,
}, func() float64 {
fetched := lists.Fetched(listURL)
if fetched.IsZero() {
return 0
}
return float64(fetched.Unix())
}),
)
}
}
// AddAbuseIPDB adds the metrics of AbuseIPDB, with the source abuseipdb,
// read from abuseIPDB as the metrics are asked for: the checks made, those
// that failed, and how many checks the day's budget has left. It is
// called once, after AddReputation, while SWW_WAF_ABUSEIPDB_KEY is set.
func (m *Metrics) AddAbuseIPDB(abuseIPDB *reputation.AbuseIPDB) {
source := prometheus.Labels{sourceLabel: reputation.AbuseIPDBSource}
m.addReputationQueries(source, abuseIPDB.Checked)
m.addReputationFailures(source, abuseIPDB.Failures)
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_reputation_daily_budget_remaining",
Help: "Checks of the day's SWW_WAF_ABUSEIPDB_DAILY_BUDGET not yet spent.",
ConstLabels: source,
}, func() float64 {
return float64(abuseIPDB.BudgetLeft())
}),
)
}
// ReputationHit counts a request whose client source lists: a blocklist,
// by its URL, a DNSBL zone, its key masked, or AbuseIPDB, abuseipdb.
func (m *Metrics) ReputationHit(source string) {
m.reputationHits.WithLabelValues(source).Inc()
}
// AddAlerts adds the metrics of the alerts sent to each destination set,
// read from queue as the metrics are asked for, by destination: the
// alerts sent, the requests to the destination that failed, the alerts
// held back, which are the same for every destination, and those
// dropped. With no destination set, it adds none.
func (m *Metrics) AddAlerts(queue *alerts.Queue) {
for _, name := range queue.DestinationsSet() {
destination := prometheus.Labels{"destination": name}
m.registry.MustRegister(
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_alerts_sent_total",
Help: "Alerts the destination took.",
ConstLabels: destination,
}, func() float64 {
return float64(queue.Counts(name).Sent)
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_alert_send_failures_total",
Help: "Requests to the destination that failed.",
ConstLabels: destination,
}, func() float64 {
return float64(queue.Counts(name).SendFailures)
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_alerts_suppressed_total",
Help: "Alerts held back: repeats within SWW_WAF_ALERT_COOLDOWN, and " +
"alerts past SWW_WAF_ALERT_MAX_PER_HOUR, for the hour's summary.",
ConstLabels: destination,
}, func() float64 {
return float64(queue.Suppressed())
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_alerts_dropped_total",
Help: "Alerts dropped, the oldest first, from a full queue, and " +
"alerts given up as the destination refused them.",
ConstLabels: destination,
}, func() float64 {
return float64(queue.Counts(name).Dropped)
}),
)
}
}
// ServeHTTP answers with the metrics in the Prometheus text format.
func (m *Metrics) ServeHTTP(w http.ResponseWriter, r *http.Request) {
m.handler.ServeHTTP(w, r)
@@ -311,12 +496,6 @@ func (m *Metrics) CoreRuleSetMatched(mode string, id int) {
m.coreRuleSetMatches.WithLabelValues(mode, strconv.Itoa(id)).Inc()
}
// ReputationHit counts a request whose client source lists: a blocklist,
// by its URL, a DNSBL zone, its key masked, or AbuseIPDB, abuseipdb.
func (m *Metrics) ReputationHit(source string) {
m.reputationHits.WithLabelValues(source).Inc()
}
// StateFileWritten counts a write of the state or cache file name, of size
// bytes, that ended with err.
func (m *Metrics) StateFileWritten(name string, size int, err error) {
@@ -348,310 +527,6 @@ func (m *Metrics) StateFileEditSetAside(name string) {
m.stateFileEditsSetAside.WithLabelValues(name).Inc()
}
// banLedger is what the metrics read from the ledger of bans.
type banLedger interface {
// Made returns the bans made since the start with cause.
Made(cause string) int
// Count returns the bans active at now, the permanent ones included,
// and the permanent ones.
Count(now time.Time) (active, permanent int)
}
// addBans adds the metrics read from ledger as the metrics are asked for:
// the bans made since the start, by cause, and the bans active and
// permanent at now.
func (m *Metrics) addBans(ledger banLedger, now func() time.Time) {
for _, cause := range []string{
bans.CauseLimit, bans.CauseAttack, bans.CauseAdmin, bans.CauseAbuseIPDB,
} {
m.registry.MustRegister(
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_bans_made_total",
Help: "Bans made, by cause.",
ConstLabels: prometheus.Labels{"cause": cause},
}, func() float64 {
return float64(ledger.Made(cause))
}),
)
}
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_active_bans",
Help: "Bans active now, the permanent ones included.",
}, func() float64 {
active, _ := ledger.Count(now())
return float64(active)
}),
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_permanent_bans",
Help: "Permanent bans not lifted.",
}, func() float64 {
_, permanent := ledger.Count(now())
return float64(permanent)
}),
)
}
// sized is what the metrics read from the table of clients and from the
// rule files.
type sized interface {
// Len returns how many clients the table holds, or how many rules were
// loaded from the rule files.
Len() int
}
// addClients adds the metric of the clients in table, read as the metrics
// are asked for.
func (m *Metrics) addClients(table sized) {
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_tracked_clients",
Help: "Clients in the table of clients.",
}, func() float64 {
return float64(table.Len())
}),
)
}
// addRules adds the metric of the rules loaded from files, read as the
// metrics are asked for.
func (m *Metrics) addRules(files sized) {
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_rules_loaded",
Help: "Rules loaded from the rule files.",
}, func() float64 {
return float64(files.Len())
}))
}
// remoteLogSender is what the metrics read from what sends the log lines to
// SWW_WAF_LOG_REMOTE_URL.
type remoteLogSender interface {
// Sent returns the lines sent.
Sent() int64
// Dropped returns the lines dropped.
Dropped() int64
// Depth returns the lines waiting in the buffer.
Depth() int
}
// addRemoteLog adds the metrics of sending the log lines to
// SWW_WAF_LOG_REMOTE_URL, read from remote as the metrics are asked for: the
// lines sent, those dropped, and those waiting in the buffer.
func (m *Metrics) addRemoteLog(remote remoteLogSender) {
m.registry.MustRegister(
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_remote_log_lines_sent_total",
Help: "Log lines sent to SWW_WAF_LOG_REMOTE_URL.",
}, func() float64 {
return float64(remote.Sent())
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_remote_log_lines_dropped_total",
Help: "Log lines dropped: the oldest in a full buffer, and those " +
"whose sending failed.",
}, func() float64 {
return float64(remote.Dropped())
}),
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_remote_log_lines_waiting",
Help: "Log lines in the buffer, waiting to be sent.",
}, func() float64 {
return float64(remote.Depth())
}),
)
}
// lookupDatabase is what the metrics read from the lookup database.
type lookupDatabase interface {
// LastRead returns when the file in use was read.
LastRead() time.Time
// ReadFailures returns the replacements of the file that could not be
// read.
ReadFailures() int
}
// addLookupFile adds the metrics of the lookup database, read from file as
// the metrics are asked for: when the file in use was read, and the
// replacements of it that could not be read.
func (m *Metrics) addLookupFile(file lookupDatabase) {
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_lookup_database_last_read_timestamp_seconds",
Help: "When the lookup database in use was read, in seconds since 1970.",
}, func() float64 {
return float64(file.LastRead().Unix())
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_lookup_database_read_failures_total",
Help: "Replacements of the lookup database that could not be read.",
}, func() float64 {
return float64(file.ReadFailures())
}),
)
}
// sourceLabel is the label of the reputation metrics: a list's URL, a
// DNSBL zone, its key masked, or abuseipdb.
const sourceLabel = "source"
// dnsblZones is what the metrics read from what asks the DNSBL zones.
type dnsblZones interface {
// Zones returns the zones asked, their keys unmasked.
Zones() []string
// Queries returns the queries made to zone.
Queries(zone string) int
// Failures returns the queries to zone that failed.
Failures(zone string) int
}
// addDNSBL adds the metrics of each zone dnsbl asks, by source, the zone,
// its key masked as config.MaskZoneKey masks it, read as the metrics are
// asked for: the queries made and those that failed.
func (m *Metrics) addDNSBL(dnsbl dnsblZones) {
for _, zone := range dnsbl.Zones() {
source := prometheus.Labels{sourceLabel: config.MaskZoneKey(zone)}
m.addReputationQueries(
source,
func() int { return dnsbl.Queries(zone) },
)
m.addReputationFailures(
source,
func() int { return dnsbl.Failures(zone) },
)
}
}
// fetchedLists is what the metrics read from the lists fetched from URLs.
type fetchedLists interface {
// URLs returns the lists' URLs.
URLs() []string
// Failures returns the fetches of the list at listURL that failed.
Failures(listURL string) int
// Fetched returns when the copy in use of the list at listURL was
// fetched, or the zero time while there is none.
Fetched(listURL string) time.Time
}
// addLists adds the metrics of each list fetched from a URL, by source, the
// list's URL, read from lists as the metrics are asked for: the fetches
// that failed and when the copy in use was fetched.
func (m *Metrics) addLists(lists fetchedLists) {
for _, listURL := range lists.URLs() {
source := prometheus.Labels{sourceLabel: listURL}
m.addReputationFailures(
source,
func() int { return lists.Failures(listURL) },
)
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_reputation_last_fetch_timestamp_seconds",
Help: "When the copy of the list in use was fetched, in seconds since " +
"1970, or 0 while there is none.",
ConstLabels: source,
}, func() float64 {
fetched := lists.Fetched(listURL)
if fetched.IsZero() {
return 0
}
return float64(fetched.Unix())
}),
)
}
}
// abuseIPDBChecker is what the metrics read from what checks clients with
// AbuseIPDB.
type abuseIPDBChecker interface {
// Checked returns the checks made.
Checked() int
// Failures returns the checks that failed.
Failures() int
// BudgetLeft returns how many checks the day's budget has left.
BudgetLeft() int
}
// addAbuseIPDB adds the metrics of AbuseIPDB, with the source abuseipdb,
// read from abuseIPDB as the metrics are asked for: the checks made, those
// that failed, and how many checks the day's budget has left.
func (m *Metrics) addAbuseIPDB(abuseIPDB abuseIPDBChecker) {
source := prometheus.Labels{sourceLabel: reputation.AbuseIPDBSource}
m.addReputationQueries(source, abuseIPDB.Checked)
m.addReputationFailures(source, abuseIPDB.Failures)
m.registry.MustRegister(
prometheus.NewGaugeFunc(prometheus.GaugeOpts{
Name: "sww_reputation_daily_budget_remaining",
Help: "Checks of the day's SWW_WAF_ABUSEIPDB_DAILY_BUDGET not yet spent.",
ConstLabels: source,
}, func() float64 {
return float64(abuseIPDB.BudgetLeft())
}),
)
}
// alertQueue is what the metrics read from the queue of alerts.
type alertQueue interface {
// DestinationsSet returns the names of the destinations set.
DestinationsSet() []string
// Counts returns the alerts sent to the destination name, the requests
// to it that failed, and the alerts dropped.
Counts(name string) alerts.Counts
// Suppressed returns the alerts held back.
Suppressed() int64
}
// addAlerts adds the metrics of the alerts sent to each destination set,
// read from queue as the metrics are asked for, by destination: the
// alerts sent, the requests to the destination that failed, the alerts
// held back, which are the same for every destination, and those
// dropped. With no destination set, it adds none.
func (m *Metrics) addAlerts(queue alertQueue) {
for _, name := range queue.DestinationsSet() {
destination := prometheus.Labels{"destination": name}
m.registry.MustRegister(
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_alerts_sent_total",
Help: "Alerts the destination took.",
ConstLabels: destination,
}, func() float64 {
return float64(queue.Counts(name).Sent)
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_alert_send_failures_total",
Help: "Requests to the destination that failed.",
ConstLabels: destination,
}, func() float64 {
return float64(queue.Counts(name).SendFailures)
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_alerts_suppressed_total",
Help: "Alerts held back: repeats within SWW_WAF_ALERT_COOLDOWN, and " +
"alerts past SWW_WAF_ALERT_MAX_PER_HOUR, for the hour's summary.",
ConstLabels: destination,
}, func() float64 {
return float64(queue.Suppressed())
}),
prometheus.NewCounterFunc(prometheus.CounterOpts{
Name: "sww_alerts_dropped_total",
Help: "Alerts dropped, the oldest first, from a full queue, and " +
"alerts given up as the destination refused them.",
ConstLabels: destination,
}, func() float64 {
return float64(queue.Counts(name).Dropped)
}),
)
}
}
// count counts a request that has ended, as RequestEnded and RequestReported
// describe.
func (m *Metrics) count(
-160
View File
@@ -1,160 +0,0 @@
package metrics_test
import (
"net/http"
"net/http/httptest"
"strings"
"testing"
"time"
"sneak.berlin/go/smallwebwaf/internal/alerts"
"sneak.berlin/go/smallwebwaf/internal/metrics"
)
const (
listURL = "https://lists.example/block.txt"
zone = "dnsbl.example"
destination = "webhook"
)
// Each stand-in answers with a number of its own, so that each metric
// shows which part it was read from.
type ledgerStandIn struct{}
func (ledgerStandIn) Made(string) int { return 1 }
func (ledgerStandIn) Count(time.Time) (int, int) { return 2, 1 }
type clientTableStandIn struct{}
func (clientTableStandIn) Len() int { return 3 }
type ruleFilesStandIn struct{}
func (ruleFilesStandIn) Len() int { return 4 }
type listsStandIn struct{}
func (listsStandIn) URLs() []string { return []string{listURL} }
func (listsStandIn) Failures(string) int { return 5 }
func (listsStandIn) Fetched(string) time.Time { return time.Time{} }
type dnsblStandIn struct{}
func (dnsblStandIn) Zones() []string { return []string{zone} }
func (dnsblStandIn) Queries(string) int { return 6 }
func (dnsblStandIn) Failures(string) int { return 0 }
type abuseIPDBStandIn struct{}
func (abuseIPDBStandIn) Checked() int { return 0 }
func (abuseIPDBStandIn) Failures() int { return 0 }
func (abuseIPDBStandIn) BudgetLeft() int { return 7 }
type alertQueueStandIn struct{}
func (alertQueueStandIn) DestinationsSet() []string {
return []string{destination}
}
func (alertQueueStandIn) Counts(string) alerts.Counts {
return alerts.Counts{Sent: 8}
}
func (alertQueueStandIn) Suppressed() int64 { return 0 }
type lookupDatabaseStandIn struct{}
func (lookupDatabaseStandIn) LastRead() time.Time { return time.Time{} }
func (lookupDatabaseStandIn) ReadFailures() int { return 9 }
type remoteLogStandIn struct{}
func (remoteLogStandIn) Sent() int64 { return 10 }
func (remoteLogStandIn) Dropped() int64 { return 0 }
func (remoteLogStandIn) Depth() int { return 0 }
func TestMetricsAreReadFromEachPart(t *testing.T) {
t.Parallel()
m := metrics.New(metrics.Params{
BusiestCount: 1,
InstanceName: "app",
Now: func() time.Time { return time.Time{} },
Bans: ledgerStandIn{},
Clients: clientTableStandIn{},
Rules: ruleFilesStandIn{},
Lists: listsStandIn{},
DNSBL: dnsblStandIn{},
AbuseIPDB: abuseIPDBStandIn{},
Alerts: alertQueueStandIn{},
LookupFile: lookupDatabaseStandIn{},
RemoteLog: remoteLogStandIn{},
})
scraped := scrape(t, m)
for _, line := range []string{
`sww_bans_made_total{cause="limit",instance="app"} 1`,
`sww_active_bans{instance="app"} 2`,
`sww_tracked_clients{instance="app"} 3`,
`sww_rules_loaded{instance="app"} 4`,
`sww_reputation_failures_total{instance="app",source="` + listURL + `"} 5`,
`sww_reputation_queries_total{instance="app",source="` + zone + `"} 6`,
`sww_reputation_daily_budget_remaining{instance="app",source="abuseipdb"} 7`,
`sww_alerts_sent_total{destination="` + destination + `",instance="app"} 8`,
`sww_lookup_database_read_failures_total{instance="app"} 9`,
`sww_remote_log_lines_sent_total{instance="app"} 10`,
} {
if !strings.Contains(scraped, "\n"+line+"\n") {
t.Errorf("metrics\n%s\nwant\n%s", scraped, line)
}
}
}
func TestPartThatIsOffAddsNoMetric(t *testing.T) {
t.Parallel()
scraped := scrape(t, metrics.New(metrics.Params{
BusiestCount: 1,
InstanceName: "app",
}))
for _, name := range []string{
"sww_bans_made_total",
"sww_tracked_clients",
"sww_rules_loaded",
"sww_reputation_",
"sww_alert",
"sww_lookup_database_",
"sww_remote_log_",
} {
if strings.Contains(scraped, name) {
t.Errorf("metrics\n%s\nshow %s", scraped, name)
}
}
}
// scrape returns the metrics m serves.
func scrape(t *testing.T, m *metrics.Metrics) string {
t.Helper()
scraped := httptest.NewRecorder()
m.ServeHTTP(
scraped,
httptest.NewRequestWithContext(
t.Context(), http.MethodGet, "/", http.NoBody,
),
)
return scraped.Body.String()
}
+48 -16
View File
@@ -110,8 +110,12 @@ type AbuseIPDB struct {
// scores are by client. Each is added as it is fetched and never moved
// up, so that the one fetched longest ago is the first dropped.
scores *simplelru.LRU[netip.Prefix, Score]
// checking are the clients whose check is under way.
checking map[netip.Prefix]bool
// checking are the clients whose check is under way, each with the
// function that ends it. stopped is true once StopChecking has been
// called, and checkGoroutines waits for the goroutine of each check.
checking map[netip.Prefix]context.CancelFunc
stopped bool
checkGoroutines sync.WaitGroup
// day is the day, in UTC, of the checks spent counts.
day civil.Date
spent int
@@ -134,7 +138,7 @@ func NewAbuseIPDB(params AbuseIPDBParams) *AbuseIPDB {
params: params,
httpClient: &http.Client{},
scores: scores,
checking: map[netip.Prefix]bool{},
checking: map[netip.Prefix]context.CancelFunc{},
}
}
@@ -144,11 +148,11 @@ func NewAbuseIPDB(params AbuseIPDBParams) *AbuseIPDB {
// addresses its request comes from. A client without one is checked in
// the background, by addr, the address its request came from, if
// offender, if it has committed an offence, unless its check is under
// way, a check failed less than failureDelay ago, or the day's checks
// have used up DailyBudget; Hit never waits for a check. The check that
// uses the budget up is logged and raised as a source_failure alert. ctx
// is the context of the client's request, and a check goes on after the
// request ends.
// way, a check failed less than failureDelay ago, the day's checks have
// used up DailyBudget, or StopChecking has been called; Hit never waits
// for a check. The check that uses the budget up is logged and raised as a
// source_failure alert. ctx is the context of the client's request, and a
// check goes on after the request ends, until StopChecking.
func (a *AbuseIPDB) Hit(
ctx context.Context, client netip.Prefix, addr netip.Addr, offender bool,
) (int64, bool) {
@@ -167,14 +171,21 @@ func (a *AbuseIPDB) Hit(
a.day, a.spent = today, 0
}
check := offender && !a.checking[client] && !now.Before(a.retryAt) &&
a.spent < a.params.DailyBudget
_, checking := a.checking[client]
check := offender && !checking && !a.stopped &&
!now.Before(a.retryAt) && a.spent < a.params.DailyBudget
if check {
a.checking[client] = true
checkCtx, end := context.WithCancel(context.WithoutCancel(ctx))
a.checking[client] = end
a.checks++
a.spent++
go a.check(context.WithoutCancel(ctx), client, addr)
a.checkGoroutines.Go(func() {
defer end()
a.check(checkCtx, client, addr)
})
}
usedUp := check && a.spent == a.params.DailyBudget
@@ -189,6 +200,24 @@ func (a *AbuseIPDB) Hit(
return 0, false
}
// StopChecking ends the checks under way, and returns once their goroutines
// have. A check it ends gives no score and is no failure: it is neither
// counted, logged nor raised as an alert, and checks are not left alone
// after it. The ended check stays spent from the day's budget. After
// StopChecking, Hit checks no client.
func (a *AbuseIPDB) StopChecking() {
a.mu.Lock()
a.stopped = true
for _, end := range a.checking {
end()
}
a.mu.Unlock()
a.checkGoroutines.Wait()
}
// Checked returns how many checks were made.
func (a *AbuseIPDB) Checked() int {
a.mu.Lock()
@@ -315,7 +344,8 @@ func (a *AbuseIPDB) RunBlacklist(ctx context.Context) {
// check checks client with AbuseIPDB by addr, one of its addresses, keeps
// the score as client's, and notes the check as no longer under way. A
// check that fails gives no score: it is counted, logged and raised as a
// source_failure alert, and no client is checked for failureDelay.
// source_failure alert, and no client is checked for failureDelay. A check
// ended as ctx ends, by StopChecking, is no failure.
func (a *AbuseIPDB) check(
ctx context.Context,
client netip.Prefix,
@@ -323,24 +353,26 @@ func (a *AbuseIPDB) check(
) {
score, err := a.ask(ctx, addr)
now := a.params.Now()
isFailure := err != nil && ctx.Err() == nil
a.mu.Lock()
delete(a.checking, client)
if err == nil {
switch {
case err == nil:
a.scores.Add(
client,
Score{Client: client, Score: score, Fetched: now},
)
} else {
case isFailure:
a.failures++
a.retryAt = now.Add(failureDelay)
}
a.mu.Unlock()
if err != nil {
if isFailure {
a.alert("checking a client with AbuseIPDB failed", err)
}
}
+49 -9
View File
@@ -394,6 +394,43 @@ func TestCheckNotAnsweredWithinTheTimeoutFailsAndHitNeverWaits(
})
}
func TestStopCheckingEndsTheCheckUnderWayAsNoFailureAndChecksNoMore(
t *testing.T,
) {
t.Parallel()
synctest.Test(t, func(t *testing.T) {
queue := newQueue()
p := abuseIPDBParams()
p.Alerts = queue
abuseIPDB := &abuseIPDBStandIn{hanging: true}
checker := newAbuseIPDB(abuseIPDB, p)
wantScore(t, checker, suspect, true, 0, false)
synctest.Wait()
// The check is ended, not waited out, and is no failure.
began := time.Now()
checker.StopChecking()
if waited := time.Since(began); waited != 0 {
t.Errorf("StopChecking returned after %s, want at once", waited)
}
wantFailures(t, checker, 0)
if got := waiting(queue); len(got) != 0 {
t.Errorf("alerts waiting %+v, want none", got)
}
// An offender without a score is then not checked.
wantScore(t, checker, other, true, 0, false)
synctest.Wait()
wantChecked(t, abuseIPDB, suspect)
})
}
func TestKeyIsSentInTheKeyHeaderAndNeverShown(t *testing.T) {
t.Parallel()
@@ -408,9 +445,12 @@ func TestKeyIsSentInTheKeyHeaderAndNeverShown(t *testing.T) {
scores: map[string]int64{suspect: 100},
}
checker := newAbuseIPDB(abuseIPDB, p)
m := metrics.New(metrics.Params{
BusiestCount: 1, InstanceName: instanceName, AbuseIPDB: checker,
})
m := metrics.New(1, "app")
m.AddReputation(
reputation.New(params()),
reputation.NewDNSBL(dnsblParams()),
)
m.AddAbuseIPDB(checker)
// One check that AbuseIPDB answers, and one it refuses with an answer
// that names the key.
@@ -463,9 +503,12 @@ func TestMetricsCountTheChecksTheFailuresAndTheBudgetLeft(t *testing.T) {
p := abuseIPDBParams()
p.DailyBudget = 5
checker := newAbuseIPDB(abuseIPDB, p)
m := metrics.New(metrics.Params{
BusiestCount: 1, InstanceName: instanceName, AbuseIPDB: checker,
})
m := metrics.New(1, "app")
m.AddReputation(
reputation.New(params()),
reputation.NewDNSBL(dnsblParams()),
)
m.AddAbuseIPDB(checker)
// One check that AbuseIPDB answers, and one that fails.
hitFrom(t, checker, suspect, true)
@@ -768,9 +811,6 @@ func wantBudgetLeft(t *testing.T, checker *reputation.AbuseIPDB, want int) {
}
}
// instanceName is the label instance of the metrics the tests build.
const instanceName = "app"
// scrapeMetrics returns the metrics m serves.
func scrapeMetrics(t *testing.T, m *metrics.Metrics) string {
t.Helper()
+44 -13
View File
@@ -82,8 +82,12 @@ type DNSBL struct {
// 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
// asking are the queries under way, each with the function that ends
// it. stopped is true once StopAsking has been called, and
// queryGoroutines waits for the goroutine of each query.
asking map[query]context.CancelFunc
stopped bool
queryGoroutines sync.WaitGroup
// 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.
@@ -120,7 +124,7 @@ func NewDNSBL(params DNSBLParams) *DNSBL {
params: params,
resolver: resolver,
verdicts: verdicts,
asking: map[query]bool{},
asking: map[query]context.CancelFunc{},
queries: map[string]int{},
failures: map[string]int{},
retryAt: map[string]time.Time{},
@@ -138,9 +142,10 @@ func (d *DNSBL) Zones() []string {
// 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.
// zone is left alone after a failure, maxQueries are under way, or
// StopAsking has been called; ListedBy never waits for a query. ctx is the
// context of the client's request, and a query goes on after the request
// ends, until StopAsking.
func (d *DNSBL) ListedBy(ctx context.Context, addr netip.Addr) []string {
d.mu.Lock()
defer d.mu.Unlock()
@@ -153,23 +158,47 @@ func (d *DNSBL) ListedBy(ctx context.Context, addr netip.Addr) []string {
q := query{zone: zone, client: addr}
kept, found := d.verdicts.Peek(q)
_, asking := d.asking[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
case !asking && !d.stopped && !now.Before(d.retryAt[zone]) &&
len(d.asking) < maxQueries:
queryCtx, end := context.WithCancel(context.WithoutCancel(ctx))
d.asking[q] = end
d.queries[zone]++
go d.ask(context.WithoutCancel(ctx), q)
d.queryGoroutines.Go(func() {
defer end()
d.ask(queryCtx, q)
})
}
}
return listedBy
}
// StopAsking ends the queries under way, and returns once their
// goroutines have. A query it ends gives no verdict and is no failure: it
// is neither counted, logged nor raised as an alert, and the zone is not
// left alone after it. After StopAsking, ListedBy asks no zone.
func (d *DNSBL) StopAsking() {
d.mu.Lock()
d.stopped = true
for _, end := range d.asking {
end()
}
d.mu.Unlock()
d.queryGoroutines.Wait()
}
// Queries returns how many queries were made to zone.
func (d *DNSBL) Queries(zone string) int {
d.mu.Lock()
@@ -238,27 +267,29 @@ func (d *DNSBL) Load(verdicts []Verdict) {
// 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.
// failureDelay. A query ended as ctx ends, by StopAsking, is no failure.
func (d *DNSBL) ask(ctx context.Context, q query) {
listed, err := d.lookUp(ctx, q)
now := d.params.Now()
isFailure := err != nil && ctx.Err() == nil
d.mu.Lock()
delete(d.asking, q)
if err == nil {
switch {
case err == nil:
d.verdicts.Add(q, Verdict{
Zone: q.zone, Client: q.client, Listed: listed, Fetched: now,
})
} else {
case isFailure:
d.failures[q.zone]++
d.retryAt[q.zone] = now.Add(failureDelay)
}
d.mu.Unlock()
if err != nil {
if isFailure {
const failed = "asking a DNSBL zone failed"
shown := config.MaskZoneKey(q.zone)
+43 -6
View File
@@ -210,6 +210,45 @@ func TestQueryNotAnsweredWithinTheTimeoutFails(t *testing.T) {
})
}
//
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestStopAskingEndsTheQueryUnderWayAsNoFailureAndAsksNoMore(
t *testing.T,
) {
synctest.Test(t, func(t *testing.T) {
queue := newQueue()
p := dnsblParams(zone)
p.Alerts = queue
resolver := &resolverStandIn{hanging: true}
dnsbl := newDNSBL(resolver, p)
wantZones(t, dnsbl, listed)
synctest.Wait()
// The query is ended, not waited out, and is no failure.
began := time.Now()
dnsbl.StopAsking()
if waited := time.Since(began); waited != 0 {
t.Errorf("StopAsking returned after %s, want at once", waited)
}
wantQueries(t, dnsbl, 1, 0)
if got := waiting(queue); len(got) != 0 {
t.Errorf("alerts waiting %+v, want none", got)
}
// A client without a verdict is then asked about in no zone.
wantZones(t, dnsbl, unlisted)
synctest.Wait()
wantQueries(t, dnsbl, 1, 0)
wantAsked(t, resolver, listedName)
waitForTheResolver()
})
}
//
//nolint:paralleltest // one at a time, as the comment at the top of this file says
func TestZoneThatFailsOrRefusesGivesNoVerdictAndIsLeftAloneForAMinute(
@@ -319,9 +358,8 @@ func TestMetricsCountEachZonesQueriesAndThoseThatFailed(t *testing.T) {
"99.2.0.192." + otherZone + ".": {rcode: serverFailure},
}}
dnsbl := newDNSBL(resolver, dnsblParams(zone, otherZone))
m := metrics.New(metrics.Params{
BusiestCount: 1, InstanceName: instanceName, DNSBL: dnsbl,
})
m := metrics.New(1, "app")
m.AddReputation(reputation.New(params()), dnsbl)
wantZones(t, dnsbl, listed)
synctest.Wait()
@@ -366,9 +404,8 @@ func TestZoneKeyIsMaskedInTheVerdictsTheFailuresAndTheMetrics(t *testing.T) {
"99.2.0.192." + keyed + ".": {addrs: []string{listing}},
"100.2.0.192." + keyed + ".": {rcode: serverFailure},
}}, p)
m := metrics.New(metrics.Params{
BusiestCount: 1, InstanceName: instanceName, DNSBL: dnsbl,
})
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.
+43 -12
View File
@@ -99,8 +99,12 @@ type Client struct {
// 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
// asking are the clients whose query is under way, each with the
// function that ends it. stopped is true once StopAsking has been
// called, and queryGoroutines waits for the goroutine of each query.
asking map[netip.Prefix]context.CancelFunc
stopped bool
queryGoroutines sync.WaitGroup
// retryAt is when routewatch may be asked again after a query failed.
retryAt time.Time
}
@@ -116,7 +120,7 @@ func New(params Params) *Client {
params: params,
httpClient: &http.Client{},
answers: answers,
asking: map[netip.Prefix]bool{},
asking: map[netip.Prefix]context.CancelFunc{},
}
}
@@ -124,8 +128,9 @@ func New(params Params) *Client {
// 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.
// failureKept ago, maxQueries are under way, or StopAsking has been called. ctx
// is the context of the client's request, and a query goes on after the request
// ends, until StopAsking.
func (c *Client) Route(
ctx context.Context, client netip.Prefix, addr netip.Addr,
) (Route, Missing) {
@@ -147,15 +152,38 @@ func (c *Client) Route(
return Route{}, RequestFailed
}
if !c.asking[client] && len(c.asking) < maxQueries {
c.asking[client] = true
_, asking := c.asking[client]
if !asking && !c.stopped && len(c.asking) < maxQueries {
queryCtx, end := context.WithCancel(context.WithoutCancel(ctx))
c.asking[client] = end
go c.ask(context.WithoutCancel(ctx), client, addr)
c.queryGoroutines.Go(func() {
defer end()
c.ask(queryCtx, client, addr)
})
}
return Route{}, NoAnswerYet
}
// StopAsking ends the queries under way, and returns once their goroutines
// have. A query it ends gives no answer and is no failure: it is not logged,
// and routewatch is not left alone after it. After StopAsking, Route asks
// routewatch nothing.
func (c *Client) StopAsking() {
c.mu.Lock()
c.stopped = true
for _, end := range c.asking {
end()
}
c.mu.Unlock()
c.queryGoroutines.Wait()
}
// Snapshot returns every answer still in use, sorted by client, as
// routewatch.json lists them.
func (c *Client) Snapshot() []Answer {
@@ -205,7 +233,8 @@ func (c *Client) Load(answers []Answer) {
// 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.
// those under way is logged. A query ended as ctx ends, by StopAsking, is no
// failure.
func (c *Client) ask(
ctx context.Context,
client netip.Prefix,
@@ -213,18 +242,20 @@ func (c *Client) ask(
) {
route, err := c.query(ctx, addr)
now := c.params.Now()
isFailure := err != nil && ctx.Err() == nil
c.mu.Lock()
delete(c.asking, client)
firstFailure := err != nil && !now.Before(c.retryAt)
firstFailure := isFailure && !now.Before(c.retryAt)
if err == nil {
switch {
case err == nil:
c.answers.Add(client, Answer{
Client: client, Netblock: route.Netblock, ASN: route.ASN, Answered: now,
})
} else {
case isFailure:
c.retryAt = now.Add(failureKept)
}
+48
View File
@@ -361,6 +361,54 @@ func TestHangingRoutewatchNeverDelaysAndFailsTheQuery(t *testing.T) {
})
}
func TestStopAskingEndsTheQueryUnderWayAsNoFailureAndAsksNoMore(
t *testing.T,
) {
t.Parallel()
// In a synctest bubble, as TestHangingRoutewatchNeverDelaysAndFailsTheQuery
// says.
synctest.Test(t, func(t *testing.T) {
f := &fixture{t: t, log: &syncBuffer{}}
f.client = routewatch.New(routewatch.Params{
URL: silentURL, Now: time.Now,
ProcessLog: slog.New(slog.NewTextHandler(f.log, nil)),
})
f.route()
synctest.Wait()
// The query is ended, not waited out.
began := time.Now()
f.client.StopAsking()
if waited := time.Since(began); waited != 0 {
t.Errorf("StopAsking returned after %s, want at once", waited)
}
// A query started now would fail within waitLimit. None is, and the
// one ended is no failure: nothing is logged, and routewatch is not
// left alone.
for range 2 {
if _, missing := f.route(); missing != routewatch.NoAnswerYet {
t.Errorf(
"Route gave %q, want %q",
missing,
routewatch.NoAnswerYet,
)
}
time.Sleep(waitLimit)
synctest.Wait()
}
if f.log.String() != "" {
t.Errorf("process log %q, want nothing", f.log.String())
}
})
}
func TestRepeatedRequestsMakeOneQueryAndTheAnswerExpires(t *testing.T) {
t.Parallel()
+13 -4
View File
@@ -870,10 +870,14 @@ func TestAnAnswerRProxyGivesItselfHasTenSecondsToBeSentAtLeast(
func TestAVisitorTooSlowToSendTheRestIsAnswered408(t *testing.T) {
t.Parallel()
// rproxy is waiting for the rest of the body well within this limit, even on
// a busy host, and answers when it runs out, well before dial's deadline.
const limit = 5 * time.Second
standIn := startWAF(t, allow)
app := startApp(t, greet)
addr, _ := startRProxy(t, standIn.URL, app.URL, map[string]string{
maxRequestTime: "3000",
maxRequestTime: strconv.FormatInt(limit.Milliseconds(), 10),
})
conn := dial(t, addr)
@@ -1262,7 +1266,9 @@ func TestAVisitorThatStopsALittleShortIsReadUntilTheRequestTimeLimit(
) {
t.Parallel()
const limit = 2 * time.Second
// The app's answer arrives well within this limit, even on a busy host, and
// the connection closes when it runs out, well before dial's deadline.
const limit = 5 * time.Second
standIn := startWAF(t, allow)
app := startApp(t, answerBeforeTheBody)
@@ -1427,17 +1433,20 @@ func TestTheRequestTimeLimitDoesNotCutAnAnswerOnceTheAppHasTheRequest(
const answer = "the end of a long answer"
// The request reaches the app well within this limit, even on a busy host.
const limit = 5 * time.Second
standIn := startWAF(t, allow)
app := startApp(t, func(w http.ResponseWriter, r *http.Request) {
_, _ = io.Copy(io.Discard, r.Body)
_ = http.NewResponseController(w).Flush()
time.Sleep(2100 * time.Millisecond)
time.Sleep(limit + 100*time.Millisecond)
_, _ = io.WriteString(w, answer)
})
addr, _ := startRProxy(t, standIn.URL, app.URL, map[string]string{
maxRequestTime: "2000",
maxRequestTime: strconv.FormatInt(limit.Milliseconds(), 10),
})
got := visit(
+54 -51
View File
@@ -1412,26 +1412,12 @@ func TestFiveFailedAsksKeepTheNextRequestFromAskingWAF(t *testing.T) {
t.Run(name, func(t *testing.T) {
t.Parallel()
var answering atomic.Bool
standIn := startWAF(
t,
func(w http.ResponseWriter, r *http.Request) {
if answering.Load() {
allow(w, r)
return
}
failure(w, r)
},
)
nextID := make(chan string, 1)
standIn := startWAF(t, failure)
// The app sends back the request ID, so the test knows which
// request rproxy kept from asking.
app := startApp(t, func(w http.ResponseWriter, r *http.Request) {
if answering.Load() {
nextID <- r.Header.Get(requestIDHeader)
}
id := r.Header.Get(requestIDHeader)
w.Header().Set(requestIDHeader, id)
greet(w, r)
})
addr, out := startRProxy(
@@ -1455,19 +1441,35 @@ func TestFiveFailedAsksKeepTheNextRequestFromAskingWAF(t *testing.T) {
t.Error("the circuit breaker opened before five asks failed")
}
answering.Store(true)
wantGreeting(t, get(t, addr, "/"))
out.waitFor(t, "the circuit breaker is open")
// Every ask fails, so a request that asks waf keeps the breaker open
// or opens it. On a busy host, more requests can ask: the five asks
// can take longer than five seconds, and an open breaker lets one
// request ask each second.
deadline := time.Now().Add(waitLimit)
// Had the next request asked waf, its question would have reached
// the stand-in before the request reached the app.
var id string
select {
case id = <-nextID:
default:
t.Fatal("the app did not get the next request")
var got answered
for !out.has("the circuit breaker is open") {
if time.Now().After(deadline) {
t.Fatal(
`rproxy did not write "the circuit breaker is open"`,
)
}
// The stand-in waits while its channel of questions is full, so
// it is emptied first.
for len(standIn.questions) > 0 {
<-standIn.questions
}
got = get(t, addr, "/")
wantGreeting(t, got)
}
// No question the stand-in got since it was last emptied came from
// the request kept from asking.
id := got.header.Get(requestIDHeader)
for len(standIn.questions) > 0 {
asked := <-standIn.questions
if asked.header.Get(requestIDHeader) == id {
@@ -1519,18 +1521,7 @@ func TestVisitorsThatGoAwayWhileWAFIsAskedLeaveTheBreakerClosed(
func TestAnswersThatCannotBePassedOnOpenTheBreaker(t *testing.T) {
t.Parallel()
var answering atomic.Bool
tooLarge := overOneMiB(dispositionDeny)
standIn := startWAF(t, func(w http.ResponseWriter, r *http.Request) {
if answering.Load() {
allow(w, r)
return
}
tooLarge(w, r)
})
standIn := startWAF(t, overOneMiB(dispositionDeny))
app := startApp(t, greet)
addr, out := startRProxy(t, standIn.URL, app.URL, map[string]string{
wafBackendTimeout: "60000",
@@ -1544,15 +1535,20 @@ func TestAnswersThatCannotBePassedOnOpenTheBreaker(t *testing.T) {
standIn.question(t)
}
answering.Store(true)
wantGreeting(t, get(t, addr, "/"))
out.waitFor(t, "the circuit breaker is open")
// A request that asks waf is answered 502, and one kept from asking gets the
// app's greeting. On a busy host, more requests can ask: the five asks can
// take longer than five seconds, and an open breaker lets one request ask
// each second.
deadline := time.Now().Add(waitLimit)
select {
case <-standIn.questions:
t.Error("waf was asked, want it not to be")
case <-time.After(10 * pollInterval):
got := get(t, addr, "/")
for got.status == http.StatusBadGateway && time.Now().Before(deadline) {
standIn.question(t)
got = get(t, addr, "/")
}
wantGreeting(t, got)
out.waitFor(t, "the circuit breaker is open")
}
func TestAPathUnderOwnPrefixNeverReachesTheApp(t *testing.T) {
@@ -1809,11 +1805,15 @@ func TestWaitForReportsWaitsForAReportBeingSent(t *testing.T) {
func TestReportOfAWebSocketCountsWhatItCarried(t *testing.T) {
t.Parallel()
// The app's answer to the upgrade arrives well within these limits, even on a
// busy host.
const limit = 5 * time.Second
standIn := startWAF(t, allow)
app := startApp(t, echoAfterUpgrade)
addr, out := startRProxy(t, standIn.URL, app.URL, map[string]string{
maxRequestTime: "2000",
maxResponseTime: "2000",
maxRequestTime: strconv.FormatInt(limit.Milliseconds(), 10),
maxResponseTime: strconv.FormatInt(limit.Milliseconds(), 10),
})
conn := dial(t, addr)
@@ -1838,7 +1838,10 @@ func TestReportOfAWebSocketCountsWhatItCarried(t *testing.T) {
}
// Past both time limits, the connection still carries what is sent.
time.Sleep(2100 * time.Millisecond)
time.Sleep(limit + 100*time.Millisecond)
// dial's deadline ran from before the upgrade and this wait.
_ = conn.SetDeadline(time.Now().Add(waitLimit))
const message = "still here\n"
+1 -1
View File
@@ -2426,7 +2426,7 @@ func midnight() time.Time {
// counts, with thresholds fill does not reach.
func newParams(dir string) state.Params {
discard := slog.New(slog.DiscardHandler)
m := metrics.New(metrics.Params{BusiestCount: 1, InstanceName: "app"})
m := metrics.New(1, "app")
queue := alerts.New(alerts.Params{
WebhookURL: &url.URL{Scheme: "https", Host: "alerts.example"},
Events: alerts.Events(),
+3 -1
View File
@@ -660,7 +660,9 @@ func TestCoreRuleSetExclusionsFileThatCannotBeUsedStopsTheStartWhileTheCoreRuleS
status := run(t.Context(), map[string]string{
listenAddr: localhost + ":0", stateDir: t.TempDir(), ruleFilesDir: t.TempDir(),
"SWW_WAF_CORE_RULE_SET_MODE": config.CoreRuleSetModeOff,
"SWW_WAF_CORE_RULE_SET_MODE": string(
config.CoreRuleSetModeOff,
),
"SWW_WAF_CORE_RULE_SET_EXCLUSIONS_FILE": path,
}, out)
if status != 1 {
+48 -31
View File
@@ -118,7 +118,8 @@ func RunWAF(ctx context.Context, params Params) int {
"settings", cfg)
// The last write follows the stop of every hook, and so the end of every
// loop that changes the state or raises an alert. A hook of the state
// loop that changes the state or raises an alert, and of every ask of
// GeoJS, the DNSBL zones, AbuseIPDB and routewatch. A hook of the state
// files would stop before the loops of the parts built ahead of them.
return run(ctx, app, failed, processLog, func() error {
err := files.WriteAll()
@@ -217,7 +218,7 @@ func newLookupFile(
lc fx.Lifecycle, cfg *config.WAF, processLog *slog.Logger,
alertQueue *alerts.Queue,
) (*lookup.File, error) {
if cfg.LookupSource != "file" {
if cfg.LookupSource != config.LookupSourceFile {
return nil, nil //nolint:nilnil // no lookup database, and no error
}
@@ -297,11 +298,13 @@ func newLists(
return lists
}
// newDNSBL returns what asks the DNSBL zones, none asked yet.
// newDNSBL returns what asks the DNSBL zones, none asked yet. At the stop it
// ends the queries under way, as reputation.DNSBL.StopAsking says.
func newDNSBL(
cfg *config.WAF, processLog *slog.Logger, alertQueue *alerts.Queue,
lc fx.Lifecycle, cfg *config.WAF, processLog *slog.Logger,
alertQueue *alerts.Queue,
) *reputation.DNSBL {
return reputation.NewDNSBL(reputation.DNSBLParams{
dnsbl := reputation.NewDNSBL(reputation.DNSBLParams{
Zones: cfg.DNSBLZones,
Resolver: cfg.DNSBLResolver,
CacheTTL: cfg.IPReputationCacheTTL,
@@ -310,12 +313,17 @@ func newDNSBL(
ProcessLog: processLog,
Alerts: alertQueue,
})
lc.Append(fx.StopHook(dnsbl.StopAsking))
return dnsbl
}
// newAbuseIPDB returns what checks clients with AbuseIPDB while
// SWW_WAF_ABUSEIPDB_KEY is set, none checked yet. Until waf stops, it
// fetches AbuseIPDB's blacklist as it is due, and bans the clients on it in
// ledger, as waf.BanAbuseIPDBBlacklist says.
// ledger, as waf.BanAbuseIPDBBlacklist says. At the stop it ends the checks
// under way, as reputation.AbuseIPDB.StopChecking says.
func newAbuseIPDB(
lc fx.Lifecycle, cfg *config.WAF, processLog *slog.Logger,
alertQueue *alerts.Queue, ledger *bans.Ledger,
@@ -341,24 +349,30 @@ func newAbuseIPDB(
})
lc.Append(background(abuseIPDB.RunBlacklist))
lc.Append(fx.StopHook(abuseIPDB.StopChecking))
return abuseIPDB
}
// newRouteWatch returns what asks routewatch about clients, or nil while
// SWW_WAF_ROUTEWATCH_URL is unset.
// SWW_WAF_ROUTEWATCH_URL is unset. At the stop it ends the queries under
// way, as routewatch.Client.StopAsking says.
func newRouteWatch(
cfg *config.WAF, processLog *slog.Logger,
lc fx.Lifecycle, cfg *config.WAF, processLog *slog.Logger,
) *routewatch.Client {
if cfg.RouteWatchURL == "" {
return nil
}
return routewatch.New(routewatch.Params{
routeWatch := routewatch.New(routewatch.Params{
URL: cfg.RouteWatchURL,
Now: utcNow,
ProcessLog: processLog,
})
lc.Append(fx.StopHook(routeWatch.StopAsking))
return routeWatch
}
// newMetrics returns the metrics, with those read from the other parts as
@@ -372,43 +386,42 @@ func newMetrics(
alertQueue *alerts.Queue, lookupFile *lookup.File,
remote *remotelog.Sender,
) *metrics.Metrics {
params := metrics.Params{
BusiestCount: cfg.MetricsBusiestCountriesAndASNs,
InstanceName: cfg.InstanceName,
Now: utcNow,
Bans: ledger,
Clients: limiter,
Rules: ruleFiles,
Lists: lists,
DNSBL: dnsbl,
Alerts: alertQueue,
m := metrics.New(cfg.MetricsBusiestCountriesAndASNs, cfg.InstanceName)
m.AddBansAndClients(ledger, limiter, utcNow)
m.AddRules(ruleFiles)
m.AddReputation(lists, dnsbl)
if cfg.AbuseIPDBKey != "" {
m.AddAbuseIPDB(abuseIPDB)
}
// A nil pointer in an interface field is not a nil interface, so a part
// that is off is left out rather than set to nil.
if cfg.AbuseIPDBKey != "" {
params.AbuseIPDB = abuseIPDB
}
m.AddAlerts(alertQueue)
if lookupFile != nil {
params.LookupFile = lookupFile
m.AddLookupFile(lookupFile.LastRead, lookupFile.ReadFailures)
}
if remote != nil {
params.RemoteLog = remote
m.AddRemoteLog(remote)
}
return metrics.New(params)
return m
}
// newGeoJS returns what looks clients up at GeoJS, none looked up yet. Each
// answer goes to the limiter and the ledger, as waf.AddLookup says, and each
// failure is handled as waf.GeoJSFailed says.
// failure is handled as waf.GeoJSFailed says. At the stop it ends the request
// under way, as geojs.GeoJS.StopAsking says.
func newGeoJS(
cfg *config.WAF, m *metrics.Metrics, limiter *ratelimit.Limiter,
ledger *bans.Ledger, alertQueue *alerts.Queue, processLog *slog.Logger,
lc fx.Lifecycle,
cfg *config.WAF,
m *metrics.Metrics,
limiter *ratelimit.Limiter,
ledger *bans.Ledger,
alertQueue *alerts.Queue,
processLog *slog.Logger,
) *geojs.GeoJS {
return geojs.New(geojs.Params{
geoJS := geojs.New(geojs.Params{
URL: geojs.URL,
Answered: func(answer geojs.Answer) {
waf.AddLookup(cfg, limiter, ledger, answer)
@@ -420,6 +433,10 @@ func newGeoJS(
},
WentWithoutAnswer: m.RequestsWithoutGeoJSAnswer.Inc,
})
lc.Append(fx.StopHook(geoJS.StopAsking))
return geoJS
}
// newWAFServer returns the server waf runs, with its parts and the time
+1 -1
View File
@@ -54,7 +54,7 @@ func (abuseIPDB) check(
"scored by AbuseIPDB at or over SWW_WAF_ABUSEIPDB_MIN_SCORE",
)
if rq.h.config.DNSBLAbuseIPDBAction != deny {
if rq.h.config.DNSBLAbuseIPDBAction != config.ReputationActionDeny {
return nil
}
+5 -4
View File
@@ -86,13 +86,14 @@ func (rq *request) countedBytes() int64 {
response, request := rq.out.bytes, rq.requestBytes
switch rq.h.config.BytesCounted {
case "response":
case config.BytesCountedResponse:
return response
case "request":
case config.BytesCountedRequest:
return request
default: // both
return response + request
case config.BytesCountedBoth:
}
return response + request
}
// banForLimit bans the client's netblock at now for a broken limit, the
+10 -9
View File
@@ -53,33 +53,34 @@ func (rq *request) limitPercentages() (percentage, percentage) {
if country == "" && !rq.skippedLookupWait {
unknown = percentage{
cfg.UnknownCountryLimitPercent,
"SWW_WAF_UNKNOWN_COUNTRY_LIMIT_PERCENT",
config.UnknownCountryLimitPercentSetting,
}
}
fetched := percentage{percent: whole}
if percent, listed := rq.h.lists.ASNLimitPercent(asn); listed {
fetched = percentage{percent, "SWW_WAF_ASN_LIMIT_PERCENT_URL"}
fetched = percentage{percent, config.ASNLimitPercentURLSetting}
}
blocklisted := percentage{percent: whole}
if rq.blocklisted && cfg.BlocklistAction == "limit" {
if rq.blocklisted &&
cfg.BlocklistAction == config.ReputationActionLimit {
blocklisted = percentage{
cfg.BlocklistLimitPercent,
"SWW_WAF_BLOCKLIST_ACTION",
config.BlocklistActionSetting,
}
}
dnsblAbuseIPDBListed := percentage{percent: whole}
if (rq.dnsblListed || rq.abuseIPDBHit) &&
cfg.DNSBLAbuseIPDBAction == "limit" {
cfg.DNSBLAbuseIPDBAction == config.ReputationActionLimit {
dnsblAbuseIPDBListed = percentage{
cfg.DNSBLAbuseIPDBLimitPercent,
"SWW_WAF_DNSBL_ABUSEIPDB_ACTION",
config.DNSBLAbuseIPDBActionSetting,
}
}
const tagSetting = "SWW_WAF_TAG_LIMIT_PERCENT"
const tagSetting = config.TagLimitPercentSetting
asnRequests := lowest(
given(cfg.TagLimitPercent, asnTag(asn), tagSetting),
@@ -107,7 +108,7 @@ func (rq *request) limitPercentages() (percentage, percentage) {
asnBytes = given(
cfg.ASNBytesLimitPercent,
asn,
"SWW_WAF_ASN_BYTES_LIMIT_PERCENT",
config.ASNBytesLimitPercentSetting,
)
}
@@ -115,7 +116,7 @@ func (rq *request) limitPercentages() (percentage, percentage) {
countryBytes = given(
cfg.CountryBytesLimitPercent,
country,
"SWW_WAF_COUNTRY_BYTES_LIMIT_PERCENT",
config.CountryBytesLimitPercentSetting,
)
}
+3 -1
View File
@@ -7,6 +7,7 @@ import (
"context"
"time"
"sneak.berlin/go/smallwebwaf/internal/config"
"sneak.berlin/go/smallwebwaf/internal/requestlog"
)
@@ -41,5 +42,6 @@ func (blocklists) check(
// client, given listedBy, the blocklists that list it: it does when one lists
// it and SWW_WAF_BLOCKLIST_ACTION is deny.
func (rq *request) refusedByBlocklists(listedBy []string) bool {
return len(listedBy) > 0 && rq.h.config.BlocklistAction == deny
return len(listedBy) > 0 && rq.h.config.BlocklistAction ==
config.ReputationActionDeny
}
+2 -2
View File
@@ -47,7 +47,7 @@ func (coreRuleSet) check(
rq.line.CoreRuleSetScore = &result.Score
for _, id := range result.RuleIDs {
rq.h.metrics.CoreRuleSetMatched(cfg.CoreRuleSetMode, id)
rq.h.metrics.CoreRuleSetMatched(string(cfg.CoreRuleSetMode), id)
}
threshold := cfg.CoreRuleSetScoreThreshold
@@ -99,7 +99,7 @@ func (rq *request) alertCoreRuleSetMatch(result coreruleset.Result) {
switch {
case rq.h.config.CoreRuleSetMode == config.CoreRuleSetModeDetect:
detail["mode"] = config.CoreRuleSetModeDetect
detail["mode"] = string(config.CoreRuleSetModeDetect)
case rq.h.config.Observe:
detail["mode"] = "observe"
}
+3 -1
View File
@@ -7,6 +7,7 @@ import (
"slices"
"time"
"sneak.berlin/go/smallwebwaf/internal/config"
"sneak.berlin/go/smallwebwaf/internal/requestlog"
)
@@ -33,7 +34,8 @@ func (countries) check(
return &verdict{
action: requestlog.ActionCountryDenied,
figures: map[string]any{
"country": country, "setting": "SWW_WAF_EXCLUSIVELY_ALLOWED_COUNTRIES",
"country": country,
"setting": config.ExclusivelyAllowedCountriesSetting,
},
}
}
+2 -1
View File
@@ -32,7 +32,8 @@ func (dnsbl) check(ctx context.Context, rq *request, _ time.Time) *verdict {
rq.addTag(config.DNSBLTagPrefix + strings.ToLower(zone))
}
if !rq.dnsblListed || rq.h.config.DNSBLAbuseIPDBAction != deny {
if !rq.dnsblListed ||
rq.h.config.DNSBLAbuseIPDBAction != config.ReputationActionDeny {
return nil
}
+3 -2
View File
@@ -49,7 +49,8 @@ func (rq *request) startLookUp(
) <-chan lookupResult {
result := make(chan lookupResult, 1)
if rq.h.config.LookupSource == "off" || !canBePlaced(rq.client) {
if rq.h.config.LookupSource == config.LookupSourceOff ||
!canBePlaced(rq.client) {
result <- lookupResult{}
return result
@@ -57,7 +58,7 @@ func (rq *request) startLookUp(
client := clientGroup(rq.h.config, rq.client)
if rq.h.config.LookupSource == "file" {
if rq.h.config.LookupSource == config.LookupSourceFile {
result <- lookupResult{
lookedUp: true, answer: rq.h.lookupFile.LookUp(client),
}
+20 -27
View File
@@ -551,11 +551,16 @@ func addCoreRuleSet(t *testing.T, params *waf.Params) {
params.CoreRuleSet = coreRuleSet
}
// addCounters adds to params the table of clients, the ledger and the
// anomaly counters, as sww builds them from params' settings.
// addCounters adds to params the metrics, the table of clients, the ledger
// and the anomaly counters, as sww builds them from params' settings. The
// metrics read the ledger, the table of clients and the rules.
func addCounters(params *waf.Params) {
cfg := params.Config
params.Metrics = metrics.New(
cfg.MetricsBusiestCountriesAndASNs,
cfg.InstanceName,
)
params.Limiter = ratelimit.New(ratelimit.Limits{
PerMinute: cfg.RateLimitPerMinute,
PerHour: cfg.RateLimitPerHour,
@@ -582,18 +587,26 @@ func addCounters(params *waf.Params) {
NamedNetblocks: cfg.AnomalyWatchedNets,
Alerts: params.Alerts,
})
params.Metrics.AddBansAndClients(
params.Ledger,
params.Limiter,
params.Now,
)
params.Metrics.AddRules(params.Rules)
}
// addOutsideSources adds to params the lists fetched from URLs, the DNSBL
// zones' verdicts, AbuseIPDB's scores from abuseIPDBCheckURL and its
// blacklist from abuseIPDBBlacklistURL, routewatch's answers while its
// setting is set, the metrics newMetrics returns, and GeoJS at geojsURL, as
// sww builds them from params' settings and the parts addCounters added.
// setting is set, and GeoJS at geojsURL, as sww builds
// them from params' settings and the parts addCounters added. The metrics
// read the lists, the zones' verdicts, and AbuseIPDB's while its key is set.
func addOutsideSources(
params *waf.Params, geojsURL, abuseIPDBCheckURL string,
) {
cfg := params.Config
limiter, ledger := params.Limiter, params.Ledger
m, limiter, ledger := params.Metrics, params.Limiter, params.Ledger
processLog, alertQueue := params.ProcessLog, params.Alerts
params.Lists = reputation.New(reputation.Params{
@@ -642,9 +655,6 @@ func addOutsideSources(
})
}
m := newMetrics(params)
params.Metrics = m
params.GeoJS = geojs.New(geojs.Params{
URL: geojsURL,
Answered: func(answer geojs.Answer) {
@@ -657,29 +667,12 @@ func addOutsideSources(
},
WentWithoutAnswer: m.RequestsWithoutGeoJSAnswer.Inc,
})
}
// newMetrics returns the metrics, as sww builds them from params' settings,
// read from the ledger, the table of clients, the rules, the lists, the
// zones' verdicts, and AbuseIPDB's while its key is set.
func newMetrics(params *waf.Params) *metrics.Metrics {
cfg := params.Config
metricsParams := metrics.Params{
BusiestCount: cfg.MetricsBusiestCountriesAndASNs,
InstanceName: cfg.InstanceName,
Now: params.Now,
Bans: params.Ledger,
Clients: params.Limiter,
Rules: params.Rules,
Lists: params.Lists,
DNSBL: params.DNSBL,
}
m.AddReputation(params.Lists, params.DNSBL)
if cfg.AbuseIPDBKey != "" {
metricsParams.AbuseIPDB = params.AbuseIPDB
m.AddAbuseIPDB(params.AbuseIPDB)
}
return metrics.New(metricsParams)
}
// newClient returns an HTTP client that sends requests as they are made,
-4
View File
@@ -7,10 +7,6 @@ import (
"sneak.berlin/go/smallwebwaf/internal/bans"
)
// deny is the SWW_WAF_BLOCKLIST_ACTION and the SWW_WAF_DNSBL_ABUSEIPDB_ACTION that
// refuses the requests of a client a source lists.
const deny = "deny"
// noteListed notes each of sources, the URLs of the blocklists or the
// DNSBL zones, their keys masked, that list the client, as noteHit does,
// with reason.
+1 -1
View File
@@ -421,7 +421,7 @@ func (rq *request) answerAtTheEnd() (geojs.Answer, bool) {
return geojs.Answer{}, false
}
if rq.h.config.LookupSource == "file" {
if rq.h.config.LookupSource == config.LookupSourceFile {
return rq.lookupAnswer, true
}