tree 65a5b7e1b56c690ee3e9347e753d484976379d4e
parent 2dec6672edfc935969869e8b1ff2d10df5f3089a
author Thomas Draebing <thomas.draebing@sap.com> 1549961193 +0100
committer Thomas Draebing <thomas.draebing@sap.com> 1553602222 +0100

Add readiness and liveness probes for gerrit-master

The kubelet did not have a way to see, whether Gerrit was running in the
gerrit-master pods created by the gerrit-master helm chart. Thus, the
readiness of the pod was reported to early, since by default just the
start of the entrypoint script is used and Gerrit startup took a few
seconds longer. Further, if Gerrit crashed, the pod would not restart,
since 'tail' was the process running in the foreground and thus used by
default to check the container health.

This change adds a readiness- and liveness probe to the gerrit-master
container in the respective pod and chart. These probes check for
a successful HTTP-connection on port 8080 of the pod, which is used to
connect to Gerrit. The initial delay and probe period can be configured
in the chart.

Change-Id: Ia2e13dbe809a4ab3168729614efb8ff064abfd2e