All

running computation Monte Carlo jobs on Google Compute Engine. Last time I ran them was September, and things have changed a bit since then. I used to run a lot of jobs with kubectl from some pod.json file, no RC, no restart, fire-and-forget setup. After I started jobs I used to get pods (kubectl get pods) and typically output looks like

NAME                READY     STATUS    RESTARTS   AGE
r8o3il08c25-y0z10   1/1       Running   0          56m
r8o3il08c25-y0z15   0/1       Pending   0          56m

After one is done and second is started, I used to get output

NAME                READY     STATUS    RESTARTS   AGE
r8o3il08c25-y0z10   1/1       Exit:0    0          1h
r8o3il08c25-y0z15   1/1       Running   0          1h

So I could, using simple grep, get the picture how many are running, how many are pending, and how many are done, and query exit code (so to check if there are errors with some pods) etc

Now output with latest SDK (Google Cloud SDK 0.9.87) looks like this

NAME      READY     STATUS    RESTARTS   AGE

All finished pods are now invisible.

Could I get old behavior back? Why it was changed?