Sidecar should have a livenessProbe per default

[johscheuer]

Currently the sidecar only has a readinessProbe to check if the Python process already has started up we should additionally add a livenessProbe so that the container get's restarted when the Python process hangs. We've seen a few Python processes in the Sidecar that have stopped working and killing the process resolves the issue. With a livenessProbe we don't have to manually kill the process. For the livenessProbe I would propose a default periodSeconds of 30s and a failureThreshold of 2 which means after 2 minutes the container will be killed if not responding. We should add a field to enable this in the cluster spec and once we move to 1.0.0 we will make this the default.