1

I am trying to find event related to my pod kuebctl describe pod <pod_name> but I am seeing blank events.

Events:          <none>

I have the application deployed in AWS EKS. I think this has started to happen when one of my nodes got replaced with another one. How do i ensure that i see the events.

when I see the output of kubectl get pods I see restart count = 1 for one of my pods which indicate there should be some events.

Any help on how to investigate this further would be really great, thanks.

Thanks.

opensource-developer
  • 2,826
  • 4
  • 38
  • 88

2 Answers2

2

It is normal to have no events on the pods if no event was generated in the last 60 minutes. I have same behavior in my cluster as well:

kubectl describe pod prometheus-77566c9987-95g92 -n istio-system | grep -i events
Events:          <none>

The default events-ttl(time to live) is 60 minutes.

Actually, while trying to decrease my ttl to reproduce an see if the events disappear without having to wait, I went into this SR , that's asking for this value to be configurable via the AWS web portal.

For longer lived and advanced logging, you need to persist the events/logs or leverage the built-in logging systems offered by your cloud provider. If you want to do it yourself, there are plenty of options for doing this(Stackdriver,Prometheues,ELK).

However, if you want to increase the ttl of the events, you must change the config through the api-server as explained in this post.

Neo Anderson
  • 5,957
  • 2
  • 12
  • 29
1

The events are only available for around 1 hr. If you perform some operation on pod then you will see events for next 1 hr.

Dashrath Mundkar
  • 7,956
  • 2
  • 28
  • 42