The output of the "failed" taskrun is not correct

[vincent-pli]

Expected Behavior

I got a failed taskrun and use kubectl describe taskrun xxxx try to found what's going on.
And it tell me that:

Warning  Failed  10m   taskrun-controller  "step-pr-source-pullrequest-4cxr4" exited with code 1 (image: "docker.io/vincentpli/pullrequest-init-4e60f6acf9725cba4c9b0c81d0ba89b8@sha256:d4e18b1212147c60c52d4189329b2553c0601f76da2282af9288fa82eeb21e89"); for logs run: kubectl -n default logs pullrequest-pod-dqfjf -c step-pr-source-pullrequest-4cxr4

That's great, but the container the message hinted is not correct, actually another container several steps before it caused the exception.

Actual Behavior

Point to a incorrect container.

Additional Info

The pod.podstatus.containerStatuses is sorted by name, the behavior cannot satisfy us, should sort by the finishedAt

[vdemeester]

/kind bug