sprint cloud data flow server kubernetes liveness timeout values

Question

I am using SCDF for kubernetes to deploy streams. Some of the kubernetes pods deployed by SCDF server are in constant restarts because the livenessProbe initialDelaySeconds of 10s is too short:

#> kubectl get pods
NAME                                READY     STATUS             RESTARTS   AGE
datapipeline-confirmation-0-g261e   0/1       CrashLoopBackOff   37         2h

As these pods are created by SCDF, I can not figure out how to tell SCDF server to use larger timeout values when deploying pods. Tried just about everything short of diving into SCDF java source code, and asking on StackExchange. Thanks in advance!

Answer 1

You could try to PATCH the pod config directly

something like:

kubectl patch pod datapipeline-confirmation-0-g261e -c <container name> -p '{"readinessProbe": {"timeoutSeconds": 60}}'

assuming you have the credentials to do so. Otherwise you might need to ask SCDF to fix their problem.