forked from cortexproject/cortex
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathruler.go
1381 lines (1190 loc) · 49.8 KB
/
ruler.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
package ruler
import (
"context"
"flag"
"fmt"
"hash/fnv"
"net/http"
"net/url"
"path/filepath"
"strings"
"sync"
"time"
"github.com/go-kit/log"
"github.com/go-kit/log/level"
"github.com/pkg/errors"
v1 "github.com/prometheus/client_golang/api/prometheus/v1"
"github.com/prometheus/client_golang/prometheus"
"github.com/prometheus/client_golang/prometheus/promauto"
"github.com/prometheus/prometheus/model/labels"
"github.com/prometheus/prometheus/model/rulefmt"
"github.com/prometheus/prometheus/notifier"
"github.com/prometheus/prometheus/promql/parser"
promRules "github.com/prometheus/prometheus/rules"
"github.com/prometheus/prometheus/util/strutil"
"github.com/weaveworks/common/user"
"golang.org/x/sync/errgroup"
"github.com/cortexproject/cortex/pkg/cortexpb"
"github.com/cortexproject/cortex/pkg/ring"
"github.com/cortexproject/cortex/pkg/ring/kv"
"github.com/cortexproject/cortex/pkg/ruler/rulespb"
"github.com/cortexproject/cortex/pkg/ruler/rulestore"
"github.com/cortexproject/cortex/pkg/tenant"
"github.com/cortexproject/cortex/pkg/util"
util_api "github.com/cortexproject/cortex/pkg/util/api"
"github.com/cortexproject/cortex/pkg/util/concurrency"
"github.com/cortexproject/cortex/pkg/util/flagext"
"github.com/cortexproject/cortex/pkg/util/grpcclient"
util_log "github.com/cortexproject/cortex/pkg/util/log"
"github.com/cortexproject/cortex/pkg/util/services"
"github.com/cortexproject/cortex/pkg/util/validation"
)
var (
supportedShardingStrategies = []string{util.ShardingStrategyDefault, util.ShardingStrategyShuffle}
// Validation errors.
errInvalidShardingStrategy = errors.New("invalid sharding strategy")
errInvalidTenantShardSize = errors.New("invalid tenant shard size, the value must be greater than 0")
errInvalidMaxConcurrentEvals = errors.New("invalid max concurrent evals, the value must be greater than 0")
)
const (
// ringKey is the key under which we store the rulers ring in the KVStore.
ringKey = "ring"
// Number of concurrent group list and group loads operations.
loadRulesConcurrency = 10
fetchRulesConcurrency = 16
rulerSyncReasonInitial = "initial"
rulerSyncReasonPeriodic = "periodic"
rulerSyncReasonRingChange = "ring-change"
// Limit errors
errMaxRuleGroupsPerUserLimitExceeded = "per-user rule groups limit (limit: %d actual: %d) exceeded"
errMaxRulesPerRuleGroupPerUserLimitExceeded = "per-user rules per rule group limit (limit: %d actual: %d) exceeded"
// errors
errListAllUser = "unable to list the ruler users"
alertingRuleFilter string = "alert"
recordingRuleFilter string = "record"
firingStateFilter string = "firing"
pendingStateFilter string = "pending"
inactiveStateFilter string = "inactive"
unknownHealthFilter string = "unknown"
okHealthFilter string = "ok"
errHealthFilter string = "err"
)
type DisabledRuleGroupErr struct {
Message string
}
func (e *DisabledRuleGroupErr) Error() string {
return e.Message
}
// Config is the configuration for the recording rules server.
type Config struct {
// This is used for template expansion in alerts; must be a valid URL.
ExternalURL flagext.URLValue `yaml:"external_url"`
// Labels to add to all alerts
ExternalLabels labels.Labels `yaml:"external_labels,omitempty" doc:"nocli|description=Labels to add to all alerts."`
// GRPC Client configuration.
ClientTLSConfig grpcclient.Config `yaml:"ruler_client"`
// How frequently to evaluate rules by default.
EvaluationInterval time.Duration `yaml:"evaluation_interval"`
// How frequently to poll for updated rules.
PollInterval time.Duration `yaml:"poll_interval"`
// Path to store rule files for prom manager.
RulePath string `yaml:"rule_path"`
// URL of the Alertmanager to send notifications to.
// If you are configuring the ruler to send to a Cortex Alertmanager,
// ensure this includes any path set in the Alertmanager external URL.
AlertmanagerURL string `yaml:"alertmanager_url"`
// Whether to use DNS SRV records to discover Alertmanager.
AlertmanagerDiscovery bool `yaml:"enable_alertmanager_discovery"`
// How long to wait between refreshing the list of Alertmanager based on DNS service discovery.
AlertmanagerRefreshInterval time.Duration `yaml:"alertmanager_refresh_interval"`
// Capacity of the queue for notifications to be sent to the Alertmanager.
NotificationQueueCapacity int `yaml:"notification_queue_capacity"`
// HTTP timeout duration when sending notifications to the Alertmanager.
NotificationTimeout time.Duration `yaml:"notification_timeout"`
// Client configs for interacting with the Alertmanager
Notifier NotifierConfig `yaml:"alertmanager_client"`
// Max time to tolerate outage for restoring "for" state of alert.
OutageTolerance time.Duration `yaml:"for_outage_tolerance"`
// Minimum duration between alert and restored "for" state. This is maintained only for alerts with configured "for" time greater than grace period.
ForGracePeriod time.Duration `yaml:"for_grace_period"`
// Minimum amount of time to wait before resending an alert to Alertmanager.
ResendDelay time.Duration `yaml:"resend_delay"`
ConcurrentEvalsEnabled bool `yaml:"concurrent_evals_enabled"`
MaxConcurrentEvals int64 `yaml:"max_concurrent_evals"`
// Enable sharding rule groups.
EnableSharding bool `yaml:"enable_sharding"`
ShardingStrategy string `yaml:"sharding_strategy"`
SearchPendingFor time.Duration `yaml:"search_pending_for"`
Ring RingConfig `yaml:"ring"`
FlushCheckPeriod time.Duration `yaml:"flush_period"`
EnableAPI bool `yaml:"enable_api"`
APIDeduplicateRules bool `yaml:"api_deduplicate_rules"`
EnabledTenants flagext.StringSliceCSV `yaml:"enabled_tenants"`
DisabledTenants flagext.StringSliceCSV `yaml:"disabled_tenants"`
RingCheckPeriod time.Duration `yaml:"-"`
// Field will be populated during runtime.
LookbackDelta time.Duration `yaml:"-"`
EnableQueryStats bool `yaml:"query_stats_enabled"`
DisableRuleGroupLabel bool `yaml:"disable_rule_group_label"`
}
// Validate config and returns error on failure
func (cfg *Config) Validate(limits validation.Limits, log log.Logger) error {
if !util.StringsContain(supportedShardingStrategies, cfg.ShardingStrategy) {
return errInvalidShardingStrategy
}
if cfg.ShardingStrategy == util.ShardingStrategyShuffle && limits.RulerTenantShardSize <= 0 {
return errInvalidTenantShardSize
}
if err := cfg.ClientTLSConfig.Validate(log); err != nil {
return errors.Wrap(err, "invalid ruler gRPC client config")
}
if cfg.ConcurrentEvalsEnabled && cfg.MaxConcurrentEvals <= 0 {
return errInvalidMaxConcurrentEvals
}
return nil
}
// RegisterFlags adds the flags required to config this to the given FlagSet
func (cfg *Config) RegisterFlags(f *flag.FlagSet) {
cfg.ClientTLSConfig.RegisterFlagsWithPrefix("ruler.client", f)
cfg.Ring.RegisterFlags(f)
cfg.Notifier.RegisterFlags(f)
// Deprecated Flags that will be maintained to avoid user disruption
//lint:ignore faillint Need to pass the global logger like this for warning on deprecated methods
flagext.DeprecatedFlag(f, "ruler.client-timeout", "This flag has been renamed to ruler.configs.client-timeout", util_log.Logger)
//lint:ignore faillint Need to pass the global logger like this for warning on deprecated methods
flagext.DeprecatedFlag(f, "ruler.group-timeout", "This flag is no longer functional.", util_log.Logger)
//lint:ignore faillint Need to pass the global logger like this for warning on deprecated methods
flagext.DeprecatedFlag(f, "ruler.num-workers", "This flag is no longer functional. For increased concurrency horizontal sharding is recommended", util_log.Logger)
//lint:ignore faillint Need to pass the global logger like this for warning on deprecated methods
flagext.DeprecatedFlag(f, "ruler.alertmanager-use-v2", "This flag is no longer functional. V1 API is deprecated and removed", util_log.Logger)
cfg.ExternalURL.URL, _ = url.Parse("") // Must be non-nil
f.Var(&cfg.ExternalURL, "ruler.external.url", "URL of alerts return path.")
f.DurationVar(&cfg.EvaluationInterval, "ruler.evaluation-interval", 1*time.Minute, "How frequently to evaluate rules")
f.DurationVar(&cfg.PollInterval, "ruler.poll-interval", 1*time.Minute, "How frequently to poll for rule changes")
f.StringVar(&cfg.AlertmanagerURL, "ruler.alertmanager-url", "", "Comma-separated list of URL(s) of the Alertmanager(s) to send notifications to. Each Alertmanager URL is treated as a separate group in the configuration. Multiple Alertmanagers in HA per group can be supported by using DNS resolution via -ruler.alertmanager-discovery.")
f.BoolVar(&cfg.AlertmanagerDiscovery, "ruler.alertmanager-discovery", false, "Use DNS SRV records to discover Alertmanager hosts.")
f.DurationVar(&cfg.AlertmanagerRefreshInterval, "ruler.alertmanager-refresh-interval", 1*time.Minute, "How long to wait between refreshing DNS resolutions of Alertmanager hosts.")
f.IntVar(&cfg.NotificationQueueCapacity, "ruler.notification-queue-capacity", 10000, "Capacity of the queue for notifications to be sent to the Alertmanager.")
f.DurationVar(&cfg.NotificationTimeout, "ruler.notification-timeout", 10*time.Second, "HTTP timeout duration when sending notifications to the Alertmanager.")
f.DurationVar(&cfg.SearchPendingFor, "ruler.search-pending-for", 5*time.Minute, "Time to spend searching for a pending ruler when shutting down.")
f.BoolVar(&cfg.EnableSharding, "ruler.enable-sharding", false, "Distribute rule evaluation using ring backend")
f.StringVar(&cfg.ShardingStrategy, "ruler.sharding-strategy", util.ShardingStrategyDefault, fmt.Sprintf("The sharding strategy to use. Supported values are: %s.", strings.Join(supportedShardingStrategies, ", ")))
f.DurationVar(&cfg.FlushCheckPeriod, "ruler.flush-period", 1*time.Minute, "Period with which to attempt to flush rule groups.")
f.StringVar(&cfg.RulePath, "ruler.rule-path", "/rules", "file path to store temporary rule files for the prometheus rule managers")
f.BoolVar(&cfg.EnableAPI, "experimental.ruler.enable-api", false, "Enable the ruler api")
f.BoolVar(&cfg.APIDeduplicateRules, "experimental.ruler.api-deduplicate-rules", false, "EXPERIMENTAL: Remove duplicate rules in the prometheus rules and alerts API response. If there are duplicate rules the rule with the latest evaluation timestamp will be kept.")
f.DurationVar(&cfg.OutageTolerance, "ruler.for-outage-tolerance", time.Hour, `Max time to tolerate outage for restoring "for" state of alert.`)
f.DurationVar(&cfg.ForGracePeriod, "ruler.for-grace-period", 10*time.Minute, `Minimum duration between alert and restored "for" state. This is maintained only for alerts with configured "for" time greater than grace period.`)
f.DurationVar(&cfg.ResendDelay, "ruler.resend-delay", time.Minute, `Minimum amount of time to wait before resending an alert to Alertmanager.`)
f.BoolVar(&cfg.ConcurrentEvalsEnabled, "ruler.concurrent-evals-enabled", false, `If enabled, rules from a single rule group can be evaluated concurrently if there is no dependency between each other. Max concurrency for each rule group is controlled via ruler.max-concurrent-evals flag.`)
f.Int64Var(&cfg.MaxConcurrentEvals, "ruler.max-concurrent-evals", 1, `Max concurrency for a single rule group to evaluate independent rules.`)
f.Var(&cfg.EnabledTenants, "ruler.enabled-tenants", "Comma separated list of tenants whose rules this ruler can evaluate. If specified, only these tenants will be handled by ruler, otherwise this ruler can process rules from all tenants. Subject to sharding.")
f.Var(&cfg.DisabledTenants, "ruler.disabled-tenants", "Comma separated list of tenants whose rules this ruler cannot evaluate. If specified, a ruler that would normally pick the specified tenant(s) for processing will ignore them instead. Subject to sharding.")
f.BoolVar(&cfg.EnableQueryStats, "ruler.query-stats-enabled", false, "Report the wall time for ruler queries to complete as a per user metric and as an info level log message.")
f.BoolVar(&cfg.DisableRuleGroupLabel, "ruler.disable-rule-group-label", false, "Disable the rule_group label on exported metrics")
cfg.RingCheckPeriod = 5 * time.Second
}
func (cfg *Config) RulesBackupEnabled() bool {
// If the replication factor is greater the 1, only the first replica is responsible for evaluating the rule,
// the rest of the replica will store the rule groups as backup only for API HA.
return cfg.Ring.ReplicationFactor > 1
}
// MultiTenantManager is the interface of interaction with a Manager that is tenant aware.
type MultiTenantManager interface {
// SyncRuleGroups is used to sync the Manager with rules from the RuleStore.
// If existing user is missing in the ruleGroups map, its ruler manager will be stopped.
SyncRuleGroups(ctx context.Context, ruleGroups map[string]rulespb.RuleGroupList)
// BackUpRuleGroups is used to store backups of rule groups owned by a different ruler instance.
BackUpRuleGroups(ctx context.Context, ruleGroups map[string]rulespb.RuleGroupList)
// GetRules fetches rules for a particular tenant (userID).
GetRules(userID string) []*promRules.Group
// GetBackupRules fetches rules for a particular tenant (userID) that the ruler stores for backup purposes
GetBackupRules(userID string) rulespb.RuleGroupList
// Stop stops all Manager components.
Stop()
// ValidateRuleGroup validates a rulegroup
ValidateRuleGroup(rulefmt.RuleGroup) []error
}
// Ruler evaluates rules.
//
// +---------------------------------------------------------------+
// | |
// | Query +-------------+ |
// | +------------------> | |
// | | | Store | |
// | | +----------------+ | |
// | | | Rules +-------------+ |
// | | | |
// | | | |
// | | | |
// | +----+-v----+ Filter +------------+ |
// | | +-----------> | |
// | | Ruler | | Ring | |
// | | <-----------+ | |
// | +-------+---+ Rules +------------+ |
// | | |
// | | |
// | | |
// | | Load +-----------------+ |
// | +--------------> | |
// | | Manager | |
// | | | |
// | +-----------------+ |
// | |
// +---------------------------------------------------------------+
type Ruler struct {
services.Service
cfg Config
lifecycler *ring.BasicLifecycler
ring *ring.Ring
store rulestore.RuleStore
manager MultiTenantManager
limits RulesLimits
subservices *services.Manager
subservicesWatcher *services.FailureWatcher
// Pool of clients used to connect to other ruler replicas.
clientsPool ClientsPool
ringCheckErrors prometheus.Counter
rulerSync *prometheus.CounterVec
ruleGroupStoreLoadDuration prometheus.Gauge
ruleGroupSyncDuration prometheus.Gauge
rulerGetRulesFailures *prometheus.CounterVec
allowedTenants *util.AllowedTenants
registry prometheus.Registerer
logger log.Logger
}
// NewRuler creates a new ruler from a distributor and chunk store.
func NewRuler(cfg Config, manager MultiTenantManager, reg prometheus.Registerer, logger log.Logger, ruleStore rulestore.RuleStore, limits RulesLimits) (*Ruler, error) {
return newRuler(cfg, manager, reg, logger, ruleStore, limits, newRulerClientPool(cfg.ClientTLSConfig, logger, reg))
}
func newRuler(cfg Config, manager MultiTenantManager, reg prometheus.Registerer, logger log.Logger, ruleStore rulestore.RuleStore, limits RulesLimits, clientPool ClientsPool) (*Ruler, error) {
ruler := &Ruler{
cfg: cfg,
store: ruleStore,
manager: manager,
registry: reg,
logger: logger,
limits: limits,
clientsPool: clientPool,
allowedTenants: util.NewAllowedTenants(cfg.EnabledTenants, cfg.DisabledTenants),
ringCheckErrors: promauto.With(reg).NewCounter(prometheus.CounterOpts{
Name: "cortex_ruler_ring_check_errors_total",
Help: "Number of errors that have occurred when checking the ring for ownership",
}),
rulerSync: promauto.With(reg).NewCounterVec(prometheus.CounterOpts{
Name: "cortex_ruler_sync_rules_total",
Help: "Total number of times the ruler sync operation triggered.",
}, []string{"reason"}),
ruleGroupStoreLoadDuration: promauto.With(reg).NewGauge(prometheus.GaugeOpts{
Name: "cortex_ruler_rule_group_load_duration_seconds",
Help: "Time taken to load rule groups from storage",
}),
ruleGroupSyncDuration: promauto.With(reg).NewGauge(prometheus.GaugeOpts{
Name: "cortex_ruler_rule_group_sync_duration_seconds",
Help: "The duration in seconds required to sync and load rule groups from storage.",
}),
rulerGetRulesFailures: promauto.With(reg).NewCounterVec(prometheus.CounterOpts{
Name: "cortex_ruler_get_rules_failure_total",
Help: "The total number of failed rules request sent to rulers in getShardedRules.",
}, []string{"ruler"}),
}
if len(cfg.EnabledTenants) > 0 {
level.Info(ruler.logger).Log("msg", "ruler using enabled users", "enabled", strings.Join(cfg.EnabledTenants, ", "))
}
if len(cfg.DisabledTenants) > 0 {
level.Info(ruler.logger).Log("msg", "ruler using disabled users", "disabled", strings.Join(cfg.DisabledTenants, ", "))
}
if cfg.EnableSharding {
ringStore, err := kv.NewClient(
cfg.Ring.KVStore,
ring.GetCodec(),
kv.RegistererWithKVName(prometheus.WrapRegistererWithPrefix("cortex_", reg), "ruler"),
logger,
)
if err != nil {
return nil, errors.Wrap(err, "create KV store client")
}
if err = enableSharding(ruler, ringStore); err != nil {
return nil, errors.Wrap(err, "setup ruler sharding ring")
}
}
ruler.Service = services.NewBasicService(ruler.starting, ruler.run, ruler.stopping)
return ruler, nil
}
func enableSharding(r *Ruler, ringStore kv.Client) error {
lifecyclerCfg, err := r.cfg.Ring.ToLifecyclerConfig(r.logger)
if err != nil {
return errors.Wrap(err, "failed to initialize ruler's lifecycler config")
}
// Define lifecycler delegates in reverse order (last to be called defined first because they're
// chained via "next delegate").
delegate := ring.BasicLifecyclerDelegate(r)
delegate = ring.NewLeaveOnStoppingDelegate(delegate, r.logger)
delegate = ring.NewTokensPersistencyDelegate(r.cfg.Ring.TokensFilePath, ring.JOINING, delegate, r.logger)
delegate = ring.NewAutoForgetDelegate(r.cfg.Ring.HeartbeatTimeout*ringAutoForgetUnhealthyPeriods, delegate, r.logger)
rulerRingName := "ruler"
r.lifecycler, err = ring.NewBasicLifecycler(lifecyclerCfg, rulerRingName, ringKey, ringStore, delegate, r.logger, prometheus.WrapRegistererWithPrefix("cortex_", r.registry))
if err != nil {
return errors.Wrap(err, "failed to initialize ruler's lifecycler")
}
r.ring, err = ring.NewWithStoreClientAndStrategy(r.cfg.Ring.ToRingConfig(), rulerRingName, ringKey, ringStore, ring.NewIgnoreUnhealthyInstancesReplicationStrategy(), prometheus.WrapRegistererWithPrefix("cortex_", r.registry), r.logger)
if err != nil {
return errors.Wrap(err, "failed to initialize ruler's ring")
}
return nil
}
func (r *Ruler) starting(ctx context.Context) error {
// If sharding is enabled, start the used subservices.
if r.cfg.EnableSharding {
var err error
if r.subservices, err = services.NewManager(r.lifecycler, r.ring, r.clientsPool); err != nil {
return errors.Wrap(err, "unable to start ruler subservices")
}
r.subservicesWatcher = services.NewFailureWatcher()
r.subservicesWatcher.WatchManager(r.subservices)
if err = services.StartManagerAndAwaitHealthy(ctx, r.subservices); err != nil {
return errors.Wrap(err, "unable to start ruler subservices")
}
}
// TODO: ideally, ruler would wait until its queryable is finished starting.
return nil
}
// Stop stops the Ruler.
// Each function of the ruler is terminated before leaving the ring
func (r *Ruler) stopping(_ error) error {
r.manager.Stop()
if r.subservices != nil {
_ = services.StopManagerAndAwaitStopped(context.Background(), r.subservices)
}
return nil
}
type sender interface {
Send(alerts ...*notifier.Alert)
}
// SendAlerts implements a rules.NotifyFunc for a Notifier.
// It filters any non-firing alerts from the input.
//
// Copied from Prometheus's main.go.
func SendAlerts(n sender, externalURL string) promRules.NotifyFunc {
return func(ctx context.Context, expr string, alerts ...*promRules.Alert) {
var res []*notifier.Alert
for _, alert := range alerts {
a := ¬ifier.Alert{
StartsAt: alert.FiredAt,
Labels: alert.Labels,
Annotations: alert.Annotations,
GeneratorURL: externalURL + strutil.TableLinkForExpression(expr),
}
if !alert.ResolvedAt.IsZero() {
a.EndsAt = alert.ResolvedAt
} else {
a.EndsAt = alert.ValidUntil
}
res = append(res, a)
}
if len(alerts) > 0 {
n.Send(res...)
}
}
}
func ruleGroupDisabled(ruleGroup *rulespb.RuleGroupDesc, disabledRuleGroupsForUser validation.DisabledRuleGroups) bool {
for _, disabledRuleGroupForUser := range disabledRuleGroupsForUser {
if ruleGroup.Namespace == disabledRuleGroupForUser.Namespace &&
ruleGroup.Name == disabledRuleGroupForUser.Name &&
ruleGroup.User == disabledRuleGroupForUser.User {
return true
}
}
return false
}
var sep = []byte("/")
func tokenForGroup(g *rulespb.RuleGroupDesc) uint32 {
ringHasher := fnv.New32a()
// Hasher never returns err.
_, _ = ringHasher.Write([]byte(g.User))
_, _ = ringHasher.Write(sep)
_, _ = ringHasher.Write([]byte(g.Namespace))
_, _ = ringHasher.Write(sep)
_, _ = ringHasher.Write([]byte(g.Name))
return ringHasher.Sum32()
}
func instanceOwnsRuleGroup(r ring.ReadRing, g *rulespb.RuleGroupDesc, disabledRuleGroups validation.DisabledRuleGroups, instanceAddr string, forBackup bool) (bool, error) {
hash := tokenForGroup(g)
rlrs, err := r.Get(hash, RingOp, nil, nil, nil)
if err != nil {
return false, errors.Wrap(err, "error reading ring to verify rule group ownership")
}
var ownsRuleGroup bool
if forBackup {
// Only the second up to the last replica are used as backup
for i := 1; i < len(rlrs.Instances); i++ {
if rlrs.Instances[i].Addr == instanceAddr {
ownsRuleGroup = true
break
}
}
} else {
// Even if the replication factor is set to a number bigger than 1, only the first ruler evaluates the rule group
ownsRuleGroup = rlrs.Instances[0].Addr == instanceAddr
}
if ownsRuleGroup && ruleGroupDisabled(g, disabledRuleGroups) {
return false, &DisabledRuleGroupErr{Message: fmt.Sprintf("rule group %s, namespace %s, user %s is disabled", g.Name, g.Namespace, g.User)}
}
return ownsRuleGroup, nil
}
func (r *Ruler) ServeHTTP(w http.ResponseWriter, req *http.Request) {
if r.cfg.EnableSharding {
r.ring.ServeHTTP(w, req)
} else {
var unshardedPage = `
<!DOCTYPE html>
<html>
<head>
<meta charset="UTF-8">
<title>Cortex Ruler Status</title>
</head>
<body>
<h1>Cortex Ruler Status</h1>
<p>Ruler running with shards disabled</p>
</body>
</html>`
util.WriteHTMLResponse(w, unshardedPage)
}
}
func (r *Ruler) run(ctx context.Context) error {
level.Info(r.logger).Log("msg", "ruler up and running")
tick := time.NewTicker(r.cfg.PollInterval)
defer tick.Stop()
var ringTickerChan <-chan time.Time
var ringLastState ring.ReplicationSet
if r.cfg.EnableSharding {
ringLastState, _ = r.ring.GetAllHealthy(RingOp)
ringTicker := time.NewTicker(util.DurationWithJitter(r.cfg.RingCheckPeriod, 0.2))
defer ringTicker.Stop()
ringTickerChan = ringTicker.C
}
r.syncRules(ctx, rulerSyncReasonInitial)
for {
select {
case <-ctx.Done():
return nil
case <-tick.C:
r.syncRules(ctx, rulerSyncReasonPeriodic)
case <-ringTickerChan:
// We ignore the error because in case of error it will return an empty
// replication set which we use to compare with the previous state.
currRingState, _ := r.ring.GetAllHealthy(RingOp)
if ring.HasReplicationSetChanged(ringLastState, currRingState) {
ringLastState = currRingState
r.syncRules(ctx, rulerSyncReasonRingChange)
}
case err := <-r.subservicesWatcher.Chan():
return errors.Wrap(err, "ruler subservice failed")
}
}
}
func (r *Ruler) syncRules(ctx context.Context, reason string) {
level.Debug(r.logger).Log("msg", "syncing rules", "reason", reason)
r.rulerSync.WithLabelValues(reason).Inc()
timer := prometheus.NewTimer(nil)
defer func() {
ruleGroupSyncDuration := timer.ObserveDuration().Seconds()
r.ruleGroupSyncDuration.Set(ruleGroupSyncDuration)
}()
loadedConfigs, backupConfigs, err := r.loadRuleGroups(ctx)
if err != nil {
return
}
// This will also delete local group files for users that are no longer in 'configs' map.
r.manager.SyncRuleGroups(ctx, loadedConfigs)
if r.cfg.RulesBackupEnabled() {
r.manager.BackUpRuleGroups(ctx, backupConfigs)
}
}
func (r *Ruler) loadRuleGroups(ctx context.Context) (map[string]rulespb.RuleGroupList, map[string]rulespb.RuleGroupList, error) {
timer := prometheus.NewTimer(nil)
defer func() {
storeLoadSeconds := timer.ObserveDuration().Seconds()
r.ruleGroupStoreLoadDuration.Set(storeLoadSeconds)
}()
ownedConfigs, backupConfigs, err := r.listRules(ctx)
if err != nil {
level.Error(r.logger).Log("msg", "unable to list rules", "err", err)
return nil, nil, err
}
loadedOwnedConfigs, err := r.store.LoadRuleGroups(ctx, ownedConfigs)
if err != nil {
level.Warn(r.logger).Log("msg", "failed to load some rules owned by this ruler", "count", len(ownedConfigs)-len(loadedOwnedConfigs), "err", err)
}
if r.cfg.RulesBackupEnabled() {
loadedBackupConfigs, err := r.store.LoadRuleGroups(ctx, backupConfigs)
if err != nil {
level.Warn(r.logger).Log("msg", "failed to load some rules backed up by this ruler", "count", len(backupConfigs)-len(loadedBackupConfigs), "err", err)
}
return loadedOwnedConfigs, loadedBackupConfigs, nil
}
return loadedOwnedConfigs, nil, nil
}
func (r *Ruler) listRules(ctx context.Context) (owned map[string]rulespb.RuleGroupList, backedUp map[string]rulespb.RuleGroupList, err error) {
switch {
case !r.cfg.EnableSharding:
owned, backedUp, err = r.listRulesNoSharding(ctx)
case r.cfg.ShardingStrategy == util.ShardingStrategyDefault:
owned, backedUp, err = r.listRulesShardingDefault(ctx)
case r.cfg.ShardingStrategy == util.ShardingStrategyShuffle:
owned, backedUp, err = r.listRulesShuffleSharding(ctx)
default:
return nil, nil, errors.New("invalid sharding configuration")
}
if err != nil {
return
}
for userID := range owned {
if !r.allowedTenants.IsAllowed(userID) {
level.Debug(r.logger).Log("msg", "ignoring rule groups for user, not allowed", "user", userID)
delete(owned, userID)
}
}
for userID := range backedUp {
if !r.allowedTenants.IsAllowed(userID) {
level.Debug(r.logger).Log("msg", "ignoring rule groups for user, not allowed", "user", userID)
delete(backedUp, userID)
}
}
return
}
func (r *Ruler) listRulesNoSharding(ctx context.Context) (map[string]rulespb.RuleGroupList, map[string]rulespb.RuleGroupList, error) {
allRuleGroups, err := r.store.ListAllRuleGroups(ctx)
if err != nil {
return nil, nil, err
}
for userID, groups := range allRuleGroups {
disabledRuleGroupsForUser := r.limits.DisabledRuleGroups(userID)
if len(disabledRuleGroupsForUser) == 0 {
continue
}
filteredGroupsForUser := rulespb.RuleGroupList{}
for _, group := range groups {
if !ruleGroupDisabled(group, disabledRuleGroupsForUser) {
filteredGroupsForUser = append(filteredGroupsForUser, group)
} else {
level.Info(r.logger).Log("msg", "rule group disabled", "name", group.Name, "namespace", group.Namespace, "user", group.User)
}
}
allRuleGroups[userID] = filteredGroupsForUser
}
return allRuleGroups, nil, nil
}
func (r *Ruler) listRulesShardingDefault(ctx context.Context) (map[string]rulespb.RuleGroupList, map[string]rulespb.RuleGroupList, error) {
configs, err := r.store.ListAllRuleGroups(ctx)
if err != nil {
return nil, nil, err
}
ownedConfigs := make(map[string]rulespb.RuleGroupList)
backedUpConfigs := make(map[string]rulespb.RuleGroupList)
for userID, groups := range configs {
owned := filterRuleGroups(userID, groups, r.limits.DisabledRuleGroups(userID), r.ring, r.lifecycler.GetInstanceAddr(), r.logger, r.ringCheckErrors)
if len(owned) > 0 {
ownedConfigs[userID] = owned
}
if r.cfg.RulesBackupEnabled() {
backup := filterBackupRuleGroups(userID, groups, r.limits.DisabledRuleGroups(userID), r.ring, r.lifecycler.GetInstanceAddr(), r.logger, r.ringCheckErrors)
if len(backup) > 0 {
backedUpConfigs[userID] = backup
}
}
}
return ownedConfigs, backedUpConfigs, nil
}
func (r *Ruler) listRulesShuffleSharding(ctx context.Context) (map[string]rulespb.RuleGroupList, map[string]rulespb.RuleGroupList, error) {
users, err := r.store.ListAllUsers(ctx)
if err != nil {
return nil, nil, errors.Wrap(err, "unable to list users of ruler")
}
// Only users in userRings will be used in the to load the rules.
userRings := map[string]ring.ReadRing{}
for _, u := range users {
if shardSize := r.limits.RulerTenantShardSize(u); shardSize > 0 {
subRing := r.ring.ShuffleShard(u, shardSize)
// Include the user only if it belongs to this ruler shard.
if subRing.HasInstance(r.lifecycler.GetInstanceID()) {
userRings[u] = subRing
}
} else {
// A shard size of 0 means shuffle sharding is disabled for this specific user.
// In that case we use the full ring so that rule groups will be sharded across all rulers.
userRings[u] = r.ring
}
}
if len(userRings) == 0 {
return nil, nil, nil
}
userCh := make(chan string, len(userRings))
for u := range userRings {
userCh <- u
}
close(userCh)
mu := sync.Mutex{}
owned := map[string]rulespb.RuleGroupList{}
backedUp := map[string]rulespb.RuleGroupList{}
concurrency := loadRulesConcurrency
if len(userRings) < concurrency {
concurrency = len(userRings)
}
g, gctx := errgroup.WithContext(ctx)
for i := 0; i < concurrency; i++ {
g.Go(func() error {
for userID := range userCh {
groups, err := r.store.ListRuleGroupsForUserAndNamespace(gctx, userID, "")
if err != nil {
return errors.Wrapf(err, "failed to fetch rule groups for user %s", userID)
}
filterOwned := filterRuleGroups(userID, groups, r.limits.DisabledRuleGroups(userID), userRings[userID], r.lifecycler.GetInstanceAddr(), r.logger, r.ringCheckErrors)
var filterBackup []*rulespb.RuleGroupDesc
if r.cfg.RulesBackupEnabled() {
filterBackup = filterBackupRuleGroups(userID, groups, r.limits.DisabledRuleGroups(userID), userRings[userID], r.lifecycler.GetInstanceAddr(), r.logger, r.ringCheckErrors)
}
if len(filterOwned) == 0 && len(filterBackup) == 0 {
continue
}
mu.Lock()
if len(filterOwned) > 0 {
owned[userID] = filterOwned
}
if len(filterBackup) > 0 {
backedUp[userID] = filterBackup
}
mu.Unlock()
}
return nil
})
}
err = g.Wait()
return owned, backedUp, err
}
// filterRuleGroups returns map of rule groups that given instance "owns" based on supplied ring.
// This function only uses User, Namespace, and Name fields of individual RuleGroups.
//
// Reason why this function is not a method on Ruler is to make sure we don't accidentally use r.ring,
// but only ring passed as parameter.
func filterRuleGroups(userID string, ruleGroups []*rulespb.RuleGroupDesc, disabledRuleGroups validation.DisabledRuleGroups, ring ring.ReadRing, instanceAddr string, log log.Logger, ringCheckErrors prometheus.Counter) []*rulespb.RuleGroupDesc {
// Prune the rule group to only contain rules that this ruler is responsible for, based on ring.
var result []*rulespb.RuleGroupDesc
for _, g := range ruleGroups {
owned, err := instanceOwnsRuleGroup(ring, g, disabledRuleGroups, instanceAddr, false)
if err != nil {
switch e := err.(type) {
case *DisabledRuleGroupErr:
level.Info(log).Log("msg", e.Message)
continue
default:
ringCheckErrors.Inc()
level.Error(log).Log("msg", "failed to check if the ruler replica owns the rule group", "user", userID, "namespace", g.Namespace, "group", g.Name, "err", err)
continue
}
}
if owned {
level.Debug(log).Log("msg", "rule group owned", "user", g.User, "namespace", g.Namespace, "name", g.Name)
result = append(result, g)
} else {
level.Debug(log).Log("msg", "rule group not owned, ignoring", "user", g.User, "namespace", g.Namespace, "name", g.Name)
}
}
return result
}
// filterBackupRuleGroups returns map of rule groups that given instance backs up based on supplied ring.
// This function only uses User, Namespace, and Name fields of individual RuleGroups.
//
// Reason why this function is not a method on Ruler is to make sure we don't accidentally use r.ring,
// but only ring passed as parameter.
func filterBackupRuleGroups(userID string, ruleGroups []*rulespb.RuleGroupDesc, disabledRuleGroups validation.DisabledRuleGroups, ring ring.ReadRing, instanceAddr string, log log.Logger, ringCheckErrors prometheus.Counter) []*rulespb.RuleGroupDesc {
var result []*rulespb.RuleGroupDesc
for _, g := range ruleGroups {
backup, err := instanceOwnsRuleGroup(ring, g, disabledRuleGroups, instanceAddr, true)
if err != nil {
switch e := err.(type) {
case *DisabledRuleGroupErr:
level.Info(log).Log("msg", e.Message)
continue
default:
ringCheckErrors.Inc()
level.Error(log).Log("msg", "failed to check if the ruler replica backs up the rule group", "user", userID, "namespace", g.Namespace, "group", g.Name, "err", err)
continue
}
}
if backup {
level.Debug(log).Log("msg", "rule group backed up", "user", g.User, "namespace", g.Namespace, "name", g.Name)
result = append(result, g)
} else {
level.Debug(log).Log("msg", "rule group not backed up, ignoring", "user", g.User, "namespace", g.Namespace, "name", g.Name)
}
}
return result
}
// GetRules retrieves the running rules from this ruler and all running rulers in the ring if
// sharding is enabled
func (r *Ruler) GetRules(ctx context.Context, rulesRequest RulesRequest) ([]*GroupStateDesc, error) {
userID, err := tenant.TenantID(ctx)
if err != nil {
return nil, fmt.Errorf("no user id found in context")
}
if r.cfg.EnableSharding {
return r.getShardedRules(ctx, userID, rulesRequest)
}
return r.getLocalRules(userID, rulesRequest, false)
}
func (r *Ruler) getLocalRules(userID string, rulesRequest RulesRequest, includeBackups bool) ([]*GroupStateDesc, error) {
groups := r.manager.GetRules(userID)
groupDescs := make([]*GroupStateDesc, 0, len(groups))
prefix := filepath.Join(r.cfg.RulePath, userID) + "/"
sliceToSet := func(values []string) map[string]struct{} {
set := make(map[string]struct{}, len(values))
for _, v := range values {
set[v] = struct{}{}
}
return set
}
ruleNameSet := sliceToSet(rulesRequest.RuleNames)
ruleGroupNameSet := sliceToSet(rulesRequest.RuleGroupNames)
fileSet := sliceToSet(rulesRequest.Files)
ruleType := rulesRequest.Type
alertState := rulesRequest.State
health := rulesRequest.Health
matcherSets, err := parseMatchersParam(rulesRequest.Matchers)
if err != nil {
return nil, errors.Wrap(err, "error parsing matcher values")
}
returnAlerts := ruleType == "" || ruleType == alertingRuleFilter
returnRecording := (ruleType == "" || ruleType == recordingRuleFilter) && alertState == ""
for _, group := range groups {
// The mapped filename is url path escaped encoded to make handling `/` characters easier
decodedNamespace, err := url.PathUnescape(strings.TrimPrefix(group.File(), prefix))
if err != nil {
return nil, errors.Wrap(err, "unable to decode rule filename")
}
if len(fileSet) > 0 {
if _, OK := fileSet[decodedNamespace]; !OK {
continue
}
}
if len(ruleGroupNameSet) > 0 {
if _, OK := ruleGroupNameSet[group.Name()]; !OK {
continue
}
}
interval := group.Interval()
groupDesc := &GroupStateDesc{
Group: &rulespb.RuleGroupDesc{
Name: group.Name(),
Namespace: string(decodedNamespace),
Interval: interval,
User: userID,
Limit: int64(group.Limit()),
},
EvaluationTimestamp: group.GetLastEvaluation(),
EvaluationDuration: group.GetEvaluationTime(),
}
for _, r := range group.Rules() {
if len(ruleNameSet) > 0 {
if _, OK := ruleNameSet[r.Name()]; !OK {
continue
}
}
if !returnByHealth(health, string(r.Health())) {
continue
}
if !matchesMatcherSets(matcherSets, r.Labels()) {
continue
}
lastError := ""
if r.LastError() != nil {
lastError = r.LastError().Error()
}
var ruleDesc *RuleStateDesc
switch rule := r.(type) {
case *promRules.AlertingRule:
if !returnAlerts {
continue
}
if !returnByState(alertState, rule.State().String()) {
continue
}
alerts := []*AlertStateDesc{}
for _, a := range rule.ActiveAlerts() {
alerts = append(alerts, &AlertStateDesc{
State: a.State.String(),
Labels: cortexpb.FromLabelsToLabelAdapters(a.Labels),
Annotations: cortexpb.FromLabelsToLabelAdapters(a.Annotations),
Value: a.Value,
ActiveAt: a.ActiveAt,
FiredAt: a.FiredAt,
ResolvedAt: a.ResolvedAt,
LastSentAt: a.LastSentAt,
ValidUntil: a.ValidUntil,
KeepFiringSince: a.KeepFiringSince,
})
}
ruleDesc = &RuleStateDesc{
Rule: &rulespb.RuleDesc{
Expr: rule.Query().String(),
Alert: rule.Name(),
For: rule.HoldDuration(),
KeepFiringFor: rule.KeepFiringFor(),
Labels: cortexpb.FromLabelsToLabelAdapters(rule.Labels()),
Annotations: cortexpb.FromLabelsToLabelAdapters(rule.Annotations()),
},
State: rule.State().String(),
Health: string(rule.Health()),
LastError: lastError,
Alerts: alerts,
EvaluationTimestamp: rule.GetEvaluationTimestamp(),
EvaluationDuration: rule.GetEvaluationDuration(),
}
case *promRules.RecordingRule:
if !returnRecording {
continue
}
ruleDesc = &RuleStateDesc{
Rule: &rulespb.RuleDesc{
Record: rule.Name(),
Expr: rule.Query().String(),
Labels: cortexpb.FromLabelsToLabelAdapters(rule.Labels()),
},
Health: string(rule.Health()),
LastError: lastError,
EvaluationTimestamp: rule.GetEvaluationTimestamp(),
EvaluationDuration: rule.GetEvaluationDuration(),
}
default:
return nil, errors.Errorf("failed to assert type of rule '%v'", rule.Name())
}