Probes for zabbixServer component #185
-
|
The default values.yaml does not include any liveness, readiness, or startup probes for the zabbix server component. I added a few so k8s knows if the component is healthy or not. startupProbe:
tcpSocket:
port: 10051
initialDelaySeconds: 30
periodSeconds: 5
timeoutSeconds: 3
failureThreshold: 12
successThreshold: 1and similar for the other probes. I'm running zabbix in HA and the active one is responding to the probes and saying it is up but the standby one is seemingly not responding to the probes so k8s keeps killing it and restarting it. Is it recommended to use probes in this set up? Does zabbix itself handle the HA health and failover? Do I need different probe settings? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
Beta Was this translation helpful? Give feedback.
Hi @freakytoad1!
This Helm chart supports livenessProbe, readinessProbe and startupProbe in templates and default values.yaml.
But you are right, we do not set the default values. It is optional.
I think it is not necessary to configure probes in HA mode. Zabbix checks the status of the zabbix-server component using the database. See this issue as an example: #183 (comment)
Your question may be better answered by Zabbix users/developers in the community or official channels. See the links below.