1517438 - Use one type of failure message for task timeouts: "Task aborted - max run time exceeded" vs. "Task timeout after X seconds. Force killing container."

Status: NEW

(Taskcluster :: Workers, enhancement)

Description

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

There are 2 kinds of error messages for task timeouts:

[taskcluster:error] Task aborted - max run time exceeded
e.g. bug 1420394

[taskcluster:error] Task timeout after 3600 seconds. Force killing container.
e.g. bug 1411358

Is there a reason why we use two different error messages? Should the occurrences be classified against different bugs? If not, please use only one error message to allow easier classification against the 1 bug. (The message without the seconds count makes it even easier.)

Comment 1

[Dustin J. Mitchell [:dustin] (he/him)]

I think they're from two different worker implementations..

Comment 2

[Geoff Brown [:gbrown]]

From my perspective it would be helpful if all task timeouts were reported to one bug. I try to monitor bug 1411358, but generally miss variants like bug 1447633.

When investigating particular instances, it is nice to see the max-run-time, so ideally, I'd like to see something like:

Task timeout after 3600 seconds. Force killing container. <<< useful diagnostic info, not reported as an error
[taskcluster:error] Task aborted - max run time exceeded <<< fixed error message for failure classification

Comment 3

[Dustin J. Mitchell [:dustin] (he/him)]

Geoff's suggestion seems good. That would be a change to docker-worker only.