Add /healthz provider for liveness probing

[tiran]

GET /healthz is a de-facto standard to check if a service is up, running and still healthy. A GET request to /healthz is used as liveness and readiness probe in Kubernetes and OpenShift. Custodia should have a path provider that supports /healthz w/o authentication (potentially limited to requests from localhost).

I suspect that /healthz won't work in the Unix socket case. Kubernetes / OpenShift only support TCP/IP checks. However the also support liveness scripts. A custodia-healthz script may do the trick.