search

cortexproject / cortex

A horizontally scalable, highly available, multi-tenant, long term Prometheus.
https://cortexmetrics.io/
Apache License 2.0
5.47k stars 799 forks source link

Querying basis on filter and getting metrics we should'nt get #5709

Open Marcusfve opened 11 months ago

Marcusfve commented 11 months ago

We had maintenance yesterday and suddenly we are seeing in our system, metrics which we are querying basis on filter, result is having data of other metrics & filters too. Not sure why query is not fetching correct data. We are querying data which is in Ingester only and we are running cortex 1.16 version. Problem started after maintenance was done and machines all the machines were restarted. One hour after the maintenance i got an alert CortexIngesterTSDBHeadCompactionFailed

This is what we got when we queried the data. Some metrics shouldnt be there. image

Im thinking maybe ingesters were affected by the force shutdown..

Right now i was able to fix this issue by cleaning ingester volume but i want to know why this happened in the first place and what i can do to prevent this issue.

If you need any more information then please, be free to ask.

yeya24 commented 11 months ago

It would be great to provide steps to reproduce the issue. Otherwise it would be hard for us to figure it out.

What was the time range you are querying? And it should only query ingester, not store gateway?

friedrichg commented 11 months ago

@Marcusfve also include the config if you can. hiding out secrets, of course.

Marcusfve commented 11 months ago

I rly dont have the steps to reproduce the issue...because it randomly happened and have no idea how to reproduce it. The time range was first 13 hours (that is the time range it sends queries to ingesters).

We are running cortex in docker swarm cluster and the config is following:

version: '3.9'

services:

  distributor:
    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "5"
    image: cortex_cortex_v1.16
    networks:
      - cortex-network
    ports:
      - "9009:9009"
      - "9095:9095"
    volumes:
      - /srv/certs:/etc/cortex/certs
    command:
      -server.http-listen-port=9009
      -server.http-tls-cert-path=/etc/cortex/certs/server.cer
      -server.http-tls-key-path=/etc/cortex/certs/server.key
      -target=distributor
      -distributor.ring.store=consul
      -distributor.ring.consul.hostname=consul-client:8500
      -distributor.ingestion-rate-limit-strategy=global
      -consul.hostname=consul-client:8500
      -distributor.extend-writes=true
      -auth.enabled=true
      -distributor.health-check-ingesters=true
      -distributor.ingestion-rate-limit=500000
      -distributor.replication-factor=3
      -distributor.shard-by-all-labels=true
      -ring.heartbeat-timeout=10m
      -validation.reject-old-samples=true
      -validation.reject-old-samples.max-age=12h
      -distributor.ring.prefix=collectors/
      -blocks-storage.backend=s3
      -blocks-storage.s3.bucket-name=cortex
      -blocks-storage.s3.endpoint=s3
      -blocks-storage.s3.access-key-id=access
      -blocks-storage.s3.secret-access-key=password
      -store.engine=blocks
      -ring.prefix=collectors/
      -blocks-storage.tsdb.block-ranges-period=2h
      -blocks-storage.tsdb.dir=/data/tsdb
      -ring.store=consul
      -distributor.ha-tracker.enable=true
      -distributor.ha-tracker.enable-for-all-users=true
      -distributor.ha-tracker.consul.hostname=consul-client:8500
      -tenant-federation.enabled=true
      -validation.max-length-label-value=4096
      -distributor.ingestion-burst-size=2000000
    deploy:
      mode: replicated
      replicas: 7

  ingester:
    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "5"
    image: cortex_cortex_v1.16
    networks:
      - cortex-network
    ports:
      - "9008:9009"
      - "9098:9095"
    volumes:
      - /srv/certs:/etc/cortex/certs
      - ingester-data:/data
    command:
      -server.http-listen-port=9009
      -server.http-tls-cert-path=/etc/cortex/certs/server.cer
      -server.http-tls-key-path=/etc/cortex/certs/server.key
      -target=ingester
      -ring.store=consul
      -consul.hostname=consul-client:8500
      -distributor.replication-factor=3
      -blocks-storage.backend=s3
      -auth.enabled=true
      -blocks-storage.s3.bucket-name=cortex
      -blocks-storage.s3.endpoint=s3
      -blocks-storage.s3.access-key-id=access
      -blocks-storage.s3.secret-access-key=password
      -store.engine=blocks
      -ring.prefix=collectors/
      -blocks-storage.tsdb.block-ranges-period=2h
      -blocks-storage.tsdb.dir=/data/tsdb
      -blocks-storage.tsdb.retention-period=96h
      -blocks-storage.tsdb.ship-interval=1m
      -distributor.health-check-ingesters=true
      -distributor.shard-by-all-labels=true
      -ingester.heartbeat-period=15s
      -ingester.join-after=0s
      -ingester.max-global-series-per-metric=950000
      -ingester.max-series-per-metric=0
      -ingester.num-tokens=512
      -ingester.unregister-on-shutdown=true
      -ring.heartbeat-timeout=10m
      -tenant-federation.enabled=true
      -validation.max-length-label-value=4096
      -ingester.max-global-series-per-user=9500000
      -ingester.max-series-per-user=0
    deploy:
      mode: replicated
      replicas: 7

  querier:
    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "5"
    image: cortex_cortex_v1.16
    networks:
      - cortex-network
    ports:
      - "9003:9009"
      - "9093:9095"
    volumes:
      - /srv/certs:/etc/cortex/certs
    command:
      -server.http-listen-port=9009
      -target=querier
      -blocks-storage.backend=s3
      -server.http-tls-cert-path=/etc/cortex/certs/server.cer
      -server.http-tls-key-path=/etc/cortex/certs/server.key
      -blocks-storage.s3.bucket-name=cortex
      -auth.enabled=true
      -blocks-storage.s3.endpoint=s3
      -blocks-storage.s3.access-key-id=access
      -blocks-storage.s3.secret-access-key=password
      -ring.store=consul
      -blocks-storage.bucket-store.sync-dir=/data/tsdb
      -consul.hostname=consul-client:8500
      -store-gateway.sharding-ring.store=consul
      -store-gateway.sharding-ring.consul.hostname=consul-client:8500
      -store-gateway.sharding-ring.replication-factor=3
      -store-gateway.sharding-enabled=true
      -distributor.replication-factor=3
      -distributor.shard-by-all-labels=true
      -distributor.health-check-ingesters=true
      -store.engine=blocks
      -blocks-storage.bucket-store.ignore-deletion-marks-delay=1h
      -querier.max-concurrent=20
      -querier.query-ingesters-within=13h
      -querier.query-store-after=12h
      -querier.worker-parallelism=4
      -store.max-query-length=4405h
      -blocks-storage.bucket-store.bucket-index.enabled=true
      -tenant-federation.enabled=true
    deploy:
      mode: replicated
      replicas: 3

  store-gateway:
    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "5"
    image: cortex_cortex_v1.16
    networks:
      - cortex-network
    ports:
      - "9004:9009"
      - "9094:9095"
    volumes:
      - /srv/certs:/etc/cortex/certs
      - store-gateway-data:/data
    command:
      -server.http-listen-port=9009
      -target=store-gateway
      -store-gateway.sharding-enabled=true
      -server.http-tls-cert-path=/etc/cortex/certs/server.cer
      -server.http-tls-key-path=/etc/cortex/certs/server.key
      -store-gateway.sharding-ring.store=consul
      -store-gateway.sharding-ring.consul.hostname=consul-client:8500
      -store-gateway.sharding-ring.replication-factor=3
      -store-gateway.sharding-strategy=default
      -blocks-storage.backend=s3
      -auth.enabled=true
      -blocks-storage.s3.bucket-name=cortex
      -blocks-storage.s3.endpoint=s3
      -blocks-storage.s3.access-key-id=access
      -blocks-storage.s3.secret-access-key=password
      -store-gateway.sharding-ring.wait-stability-min-duration=0
      -blocks-storage.bucket-store.sync-dir=/data/tsdb
      -store.engine=blocks
      -server.grpc.keepalive.ping-without-stream-allowed=true
      -server.grpc.keepalive.min-time-between-pings=10m
      -blocks-storage.bucket-store.ignore-deletion-marks-delay=1h
      -blocks-storage.bucket-store.bucket-index.enabled=true
      -blocks-storage.bucket-store.index-header-lazy-loading-enabled=true
      -tenant-federation.enabled=true
    deploy:
      mode: replicated
      replicas: 3

  ruler:
    image: cortex_cortex_v1.16
    networks:
      - cortex-network
    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "5"
    ports:
      - "9001:9009"
      - "9091:9095"
    volumes:
      - /srv/certs:/etc/cortex/certs
    command:
      -server.http-listen-port=9009
      -target=ruler
      -api.response-compression-enabled=true
      -server.http-tls-cert-path=/etc/cortex/certs/server.cer
      -server.http-tls-key-path=/etc/cortex/certs/server.key
      -blocks-storage.backend=s3
      -auth.enabled=true
      -blocks-storage.s3.bucket-name=cortex
      -blocks-storage.s3.endpoint=s3
      -blocks-storage.s3.access-key-id=access
      -blocks-storage.s3.secret-access-key=password
      -store.engine=blocks
      -blocks-storage.bucket-store.sync-dir=/data/tsdb
      -ring.store=consul
      -consul.hostname=consul-client:8500
      -distributor.extend-writes=true
      -distributor.health-check-ingesters=true
      -distributor.replication-factor=3
      -distributor.shard-by-all-labels=true
      -experimental.ruler.enable-api=true
      -querier.query-ingesters-within=13h
      -querier.query-store-after=12h
      -ruler-storage.backend=s3
      -ruler-storage.s3.bucket-name=ruler
      -ruler-storage.s3.endpoint=s3
      -ruler-storage.s3.secret-access-key=password
      -ruler-storage.s3.access-key-id=access
      -ruler.enable-sharding=true
      -ruler.ring.store=consul
      -ruler.ring.consul.hostname=consul-client:8500
      -store-gateway.sharding-enabled=true
      -store-gateway.sharding-ring.consul.hostname=consul-client:8500
      -store-gateway.sharding-ring.replication-factor=3
      -store-gateway.sharding-ring.store=consul
      -ruler.alertmanager-url=https://alertmanager.tst.tehik.ee/alertmanager
      -store.engine=blocks
      -tenant-federation.enabled=true
    deploy:
      mode: replicated
      replicas: 3

  alertmanager:
    image: cortex_cortex_v1.16
    networks:
      - cortex-network
    ports:
      - "9000:9009"
      - "9090:9095"

    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "5"
    volumes:
      - /srv/certs:/etc/cortex/certs
      - alertmanager-data:/data
      - /srv/alertmanager/:/etc/cortex/alertmanager/
    command:
      -server.http-listen-port=9009
      -target=alertmanager
      -auth.enabled=true
      -alertmanager-storage.backend=s3
      -server.http-tls-cert-path=/etc/cortex/certs/server.cer
      -server.http-tls-key-path=/etc/cortex/certs/server.key
      -alertmanager-storage.s3.secret-access-key=password
      -alertmanager-storage.s3.access-key-id=access
      -alertmanager-storage.s3.endpoint=s3
      -alertmanager-storage.s3.bucket-name=alertmanager-test
      -alertmanager.storage.path=/data
      -alertmanager.web.external-url=/alertmanager
      -alertmanager.configs.fallback=/etc/cortex/alertmanager/alertmanager.yml
      -experimental.alertmanager.enable-api=true
      -alertmanager.sharding-enabled=true
      -alertmanager.sharding-ring.store=consul
      -alertmanager.sharding-ring.consul.hostname=consul-client:8500
      -tenant-federation.enabled=true
    deploy:
      mode: replicated
      replicas: 3

  compactor:
    logging:
      driver: "json-file"
      options:
        max-size: "50m"
        max-file: "5"
    image: cortex_cortex_v1.16
    networks:
      - cortex-network
    ports:
      - "9006:9009"
      - "9096:9095"
    volumes:
      - /srv/certs:/etc/cortex/certs
      - compactor-data:/data
    command:
      -server.http-listen-port=9009
      -target=compactor
      -blocks-storage.backend=s3
      -server.http-tls-cert-path=/etc/cortex/certs/server.cer
      -server.http-tls-key-path=/etc/cortex/certs/server.key
      -blocks-storage.s3.bucket-name=cortex
      -auth.enabled=true
      -blocks-storage.s3.endpoint=s3
      -blocks-storage.s3.access-key-id=access
      -blocks-storage.s3.secret-access-key=password
      -compactor.sharding-enabled=true
      -compactor.ring.store=consul
      -compactor.ring.prefix=collectors/
      -compactor.ring.consul.hostname=consul-client:8500
      -store.engine=blocks
      -compactor.data-dir=/data
      -compactor.block-ranges=2h,12h,24h
      -compactor.blocks-retention-period=90d
      -compactor.cleanup-interval=15m
      -compactor.compaction-concurrency=1
      -compactor.compaction-interval=5m
      -tenant-federation.enabled=true
      -compactor.cleanup-concurrency=20
    deploy:
      mode: replicated
      replicas: 7

networks:
  cortex-network:
    name: cortex-network
    attachable: true

volumes:
  ingester-data:
  store-gateway-data:
  compactor-data:
  alertmanager-data:

Please if you need more information, just ask. And i would rly appreciate if you take a look on my config as overall and make suggestions to make something better, maybe i have missed something important.

alanprot commented 11 months ago

Do you still have the bad TSDBs? If you still have can u try to restore it on another ingester and see if the problem is still there? It would nice to have a peek on those TSDBs.

Marcusfve commented 11 months ago

I do not have the bad TSDBs anymore, but as i said i was able to fix this issue by cleaning ingester volume, maybe this can give you some hint.