Netflix/Priam

Priam is not checking the C* process health.

Opened this issue · 0 comments

cthumuluru commented

Describe the bug
Priam only checks the availability of C* process to mark C* as live or not. When C* process is in unhealthy state, although JMX probes fail continuously, Priam doesn't mark the node as unhealthy. Enabling the health-check to mark C* live or not would be really helpful. Reference: CassandraMonitor class.

To Reproduce
Steps to reproduce the behavior:

  1. Log in to one of the nodes running C* and Priam.
  2. Stop the process by passing STOP signal. kill -STOP <pid>.
  3. Priam's JMX probes keep failing and the C* process is reported as healthy.

Expected behavior
C* instance should be marked unhealthy since the JMX probes are failing.

Version
trunk