Kubernetes deployed Elastic Agent is looped in beat restart

[Alphayeeeet (Alphabeet)]

Hello,

I tried to deploy Fleet-managed Elastic Agent with Kubernetes Integration into RedHat OpenShift. When deploying an agent policy update using Fleet, the beats are caught up in a restart loop:

Agent version: beats/elastic-agent:8.11.2

After restarting the agent pods manually , the problem is solved.

This might be a bug in Elastic Agent.

I would appreciate any advice. Thank you.

[Alphayeeeet (Alphabeet)]

Reposted: Kubernetes deployed Elastic Agent is looped in beat restart

[stephenb (Stephen Brown)]

Hi @Alphayeeeet

Welcome to the community.

This is often caused by an out-of-memory issue. You might want to bump up the resources for the agent.

Also, I am going to close the other Topic as this is a fine / close enough category.

[Alphayeeeet (Alphabeet)]

Hi @stephenb

Thank you for the reply. After increasing the memory limits, the agent is running smoothly. Maybe this could be added in the troubleshooting documentation of Elastic Agent?

[stephenb (Stephen Brown)]

Actually there is a open issue to make this much easier to identify.

Report the agent status as degraded when previously OOMKilled on Kubernetes