You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
# the livenessProbe just checks to see that the admin api is listening and returning 200s.
livenessProbe:
exec:
command:
- /bin/sh
- -c
- curl --silent --fail -k "http://${SERVICE_NAME}.redpanda-cluster.prod.svc.cluster.local.:9644/v1/status/ready"
initialDelaySeconds: 10
failureThreshold: 3
periodSeconds: 10
This would avoid a slow exec out to curl and reduce the surface area of failure should a DNS resolution not work out per the way we currently do the liveness probe (k8s cluster DNS failure could cause cascading failure of the Redpanda cluster even though RP may be alive and healthy).
The text was updated successfully, but these errors were encountered:
hcoyote
changed the title
Improve liveness probe to utilize to use builtin http probe instead of curl
Improve liveness probe to use builtin http probe instead of curl
May 13, 2024
What would you like to be added?
Update the liveness probe to utilize the builtin-in http check rather than a curl.
https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-startup-probes/#define-a-liveness-http-request
See additional details in slack thread: https://redpandadata.slack.com/archives/C01H6JRQX1S/p1715624674945489
Why is this needed?
This would avoid a slow
exec
out tocurl
and reduce the surface area of failure should a DNS resolution not work out per the way we currently do the liveness probe (k8s cluster DNS failure could cause cascading failure of the Redpanda cluster even though RP may be alive and healthy).The text was updated successfully, but these errors were encountered: