Revert default runner script timeout
What does this MR do?
This reverts merge request !4491 (merged)
In addition, it now links to documentation whenever a job/script timeout is exceeded so that users know how they can better control the individual parts of the job.
Why was this MR needed?
!4491 (merged) introduced both a bug fix and a breaking change.
Many customers expect after_script
to execute on timeouts, but by enforcing this by default, jobs end earlier, which is a breaking change.
What's the best way to test this MR?
# check includes: "step_script could not run to completion"
# check failure reason is: job_execution_timeout
timeout/job_timeout_during_script:
script:
- sleep 3600
timeout: 30 seconds
# check includes: "after_script could not run to completion"
# check excludes: "after_script failed, but job will continue unaffected"
# check failure reason is: job_execution_timeout
timeout/job_timeout_during_after_script_and_job_is_unsuccessful:
script:
- echo hi
after_script:
- sleep 3600
timeout: 30 seconds
# check includes: "step_script could not run to completion"
# check failure reason is: job_execution_timeout
timeout/script_timeout:
variables:
RUNNER_SCRIPT_TIMEOUT: 10s
script:
- sleep 20
# check includes: "after_script could not run to completion"
# check includes: "after_script failed, but job will continue unaffected"
timeout/after_script_timeout_but_job_is_successful:
variables:
RUNNER_AFTER_SCRIPT_TIMEOUT: 10s
script:
- echo hi
after_script:
- sleep 3600
What are the relevant issue numbers?
Closes #37363 (closed)
Edited by Arran Walker