The current helm chart defines liveness and startup probes for the check agent. However, the check agent doesn't actually expose anything on port 80, so the monitors constantly fail. While removing the checks would solve the constant restarting issue, it'd be nice if there were actual liveness probes.
