OpenShift - Pods auto-restart

[Malaik]

Hello Members,

We started monitoring Openshift in our ENV,

a lot of pods are recycled and killed ...

I notice that each new pod (that contain the application services ...) need restart,  is there any way to monitor directly all those services, processes, or I must ask everytime the application team to restart the new pods.

What is the best practice for that.

Thanks for your help and good day.

[adam_gardner · Accepted solution]

The processes (pods) will appear automatically as you've already discovered, but code level visibility does rely on the restart. 

[Malaik]

Thanks @adam_gardner 
We need a restart every time or just the first as usual?

[adam_gardner · Accepted solution]

For the purposes of this discussion, you can effectively ignore that they're even running in pods. So if this software was running on a JVM (not in a container)... If you would traditionally require a restart for whatever you're trying to do, you'll need a pod recycle.

I don't follow what you mean by "we need a restart every time or just the first as usual" But if I was to guess, you mean for agent upgrades. In which case, it will act like any non-containerised application:

After the first restart, you'll get the agent in the pod running v1. v2 OneAgent comes along and only takes effect when the pod is "naturally" restarted (meaning Dynatrace doesn't proactively recycle pods). So you'll stay on v2 until your pods restart.

Hope that helps and if I've misinterpreted your intention, reply with more info and I'll try to answer