Hello, We have some use case where sometimes jobs are stuck in reconcile state or restarting infinitely. Do you know in that case how can we monitor and alert?

Check this metrics: flink_jobmanager_job_numRestarts and https://nightlies.apache.org/flink/flink-docs-master/docs/ops/metrics/#availability

But this is only available if you have jobmanager running, you dont get these metrics if job manager hasnt been deployed after application upgrade.

if you have no job manager you can check if the pod exists in the metrics

Yeah that is how now we have created. Looking for job / taskmanager pods