Hey, I have a question. In Flink Operator for kubernetes, if task manager failed it restarts it, but ignores readiness check and starts processing data, which in my case leads to fail loop. Any idea how to force Flink Operator to respect readiness checks?

Could you specify which readiness checks do you mean? Are you using the native or standalone kubernetes integration?

@Marton Balassi, I mean default readiness checks in kubernetes, like that:

readinessProbe:
  tcpSocket:
    port: 6123
  initialDelaySeconds: 180
  periodSeconds: 60

Version of Flink operator: 1.3.0 Not sure about that native\standalone and where to check. So probably it's default.