A readiness probe will ensure traffic won't be sent to a non ready container. But if that container is in a bad state such that it will never be "ready", it's just wasting resources (and if it's your only container (replica count =1) then your application will be unavailable and stay that way).
A failing liveness probe will cause a restart of a badly behaving container, which (hopefully) will get it back to a state where it can pass a readiness probe and begins serving traffic again.
So yes, you likely still want a liveness probe.
Ben Parees | OpenShift
Hi,I'm not 100% sure if I need to add a liveness probe when I already have a readiness probe (on a http port) configured. Can someone enlighten me on this please?Also, weird thing: If I edit a DC to add a probe (live or ready), I get a:"Edit cancelled, no changes made."message, and the dc is indeed not changed.Thanks
users mailing list
users lists openshift redhat com