CODEWITHSUNDEEP
kubernetesresiliency
Juliatzin
Juliatzin

Reputation: 19725

How to debug OOMKilled issues with Kubernetes / Why is traefik pod down

I'm testing a local app that use a service that wrap an external soap service.

When I make a request to get data from 1 week, everything works fine.

But when I make a request to get data from 1 month, my web server / reverse proxy falls ( traefik )

When I check the pods, I get:

SGE:

Last State:     Terminated
  Reason:       OOMKilled
  Exit Code:    137
  Started:      Wed, 26 Feb 2020 09:54:29 +0100
  Finished:     Wed, 26 Feb 2020 15:48:13 +0100

Traefik: 

    Last State:     Terminated
      Reason:       OOMKilled
      Exit Code:    137
      Started:      Thu, 13 Feb 2020 15:53:10 +0100
      Finished:     Wed, 26 Feb 2020 15:48:13 +0100

I have checked the Memory / CPU usage of each container with Grafana. I can't find any sign of anormal activity.

Here are the screenshot for memory Usage:

Traefik

SOAP Wrapper API

In the configuration, traefik has no limit on resources. SGE service has those limits:

      resources:
            # keep request = limit to keep this container in guaranteed class
            limits:
              cpu: 500m
              memory: 400Mi
            requests:
              cpu: 200m
              memory: 200Mi

My VM has 800 MB free memory when Idle, so it shouldn't be an issue.

Why is Traefik falling ? I don't understand what is happening, and how should I prevent it.

Any idea ???

Upvotes: 0

Views: 1658

Answers (0)

Related Questions