Google Kubernetes Engine monitoring is reset on redeployment

6/19/2019

I have two GKE projects (test and prod) where we deploy the same containers using Helm charts. Both projects use Stackdriver for Logging & Monitoring.

On our test project, I am able to see Stackdriver Monitoring charts across multiple deployments. However, on the production account, I noticed that every time I re-deploy my code, the monitoring overview tab is reset.

See https://nimb.ws/fuhTRt - As you can see the monitor starts at roughly 6 AM, which corresponds to the most recent ReplicaSet deployment at Jun 19, 2019, 5:53:44 AM

This happens even though the deployment workload does not change, only the ReplicaSet and underlying pods change.

How can I persist the monitoring metrics across deployments? - I have very little to go on!

  1. Is it some kind of project-level setting that I am mixing up?
  2. Can I provide any more information for debugging this?
-- rtindru
google-kubernetes-engine
kubernetes
monitoring
stackdriver

0 Answers