Cannot collect logs from kubernetes with containerd runtimes

Hi everyone
We deployed elastic agents with kubernetes integrations on newly installed kubernetes clusters 1.24. We noticed some missing logs and investigated the agents logs and status: we have this kind of error for every log file

11:10:17.345
elastic_agent.filebeat
[elastic_agent.filebeat][debug] Incoming log.file.path value: /var/log/containers/elastic-agent-f5tqx_monitoring_elastic-agent-c38bb0fa3843aed8800b5bbe230984693f6b1860f4e46c6910c46a4fcbd6dbb2.log
11:10:17.345
elastic_agent.filebeat
[elastic_agent.filebeat][debug] log.file.path value does not contain matcher's logs_path '/var/lib/docker/containers/', skipping...
11:10:17.345
elastic_agent.filebeat
[elastic_agent.filebeat][debug] No container match string, not adding kubernetes data

and the detailed elastic-agent status --output yaml command returns a lot of this. one for each log file

  - unit_id: filestream-default-filestream-container-logs-00222d86-dbb3-401d-8d4b-c00aeddc4fc1-kubernetes-40aada03-ef38-4dde-a6e9-68e1d06dd132.api-server
    unit_type: 0
    state: 4
    message: 'Failed: pid ''11842'' exited with code ''-1'''

Since kubernetes 1.24 uses containerd and not docker anymore, how should we configure the kubernetes integration? are those filebeat crashes related to the missing /var/lib/docker folder?

I tried every combination of processor configuration but seems it's been overridden by the main filebeat configuration which I have not accesso to, being within elastic agent.
can you help me please?

Hi @venturieffect,
Is this the same issue as Elastic Agent 8.6.x standalone deployment in Kubernetes doesn't start monitoring new pods until agent restart · Issue #2269 · elastic/elastic-agent · GitHub?

Hi Michel
It may be the case, yes.
We noticed two separate issues:

  • On one kubernetes cluster the filestream units are constantly crashing as showed in my opening post
  • on basically all other kubernetes clusters the filestream units are healty but the count is way less than the actual number of pods in the node. I also experienced what your github issue says: if the pods are restarted, they now appears among the filestream units

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.