Description of problem: The events list used to list output from lifecycle hooks (eg readiness and lifecycle probes). Without these messages from the lifecycle hooks it can be debugging certain situations very difficult. It also means we can't provide messages to users as to why the hook failed.
@mwringe are you referring to deployment hooks (pre/mid/post) and that they are not reporting errors to events OR you are referring to liveness and readiness probes not being reported in events?
liveness and readiness probes, we used to get any output that these probes produced.
Please don't open bugs in both GitHub and here. Since you have this one, I'm going to close the GitHub issue (https://github.com/openshift/origin/issues/10424). Upstream PR: https://github.com/kubernetes/kubernetes/pull/30731
This is specific to exec probes. There is a regression in Kube 1.3 that makes it so the output from exec probes is not included in the events.
Test on v3.3.0.24-dirty, "Readiness" event shows correctly steps: 1. Create a pod with readiness exec probe $ oc create -f https://raw.githubusercontent.com/mdshuai/testfile-openshift/master/k8s/readinessProbe-exec.yaml 2. Check event shows correctly. [root@openshift-107 ~]# oc get event -n dma|grep "Readiness probe failed" 4s 4m 25 nginx Pod spec.containers{nginx} Warning Unhealthy {kubelet openshift-159.lab.sjc.redhat.com} Readiness probe failed: Get http://10.1.0.6:112/: dial tcp 10.1.0.6:112: getsockopt: connection refused
[root@openshift-107 ~]# oc get event -n dma|grep "Readiness probe" 7s 37s 4 hello-pod Pod spec.containers{hello-pod} Warning Unhealthy {kubelet openshift-159.lab.sjc.redhat.com} Readiness probe failed: exec: "/bin/hello": stat /bin/hello: no such file or directory