Description
The main scale-up path has a config flag, ENABLE_JOB_QUEUED_CHECK, that skips an API call to GitHub that checks whether a job is still queued before creating a runner for it. The retry lambda does not check this flag — it always makes that call, on every retry attempt.
Impact
Turning off ENABLE_JOB_QUEUED_CHECK does not fully turn off the check: the retry path keeps calling GitHub's API for every retry, using up rate-limit budget the flag was meant to save. This makes the flag behave inconsistently depending on which lambda handles a given job.
Proposed fix
Read the same ENABLE_JOB_QUEUED_CHECK flag in the retry lambda, and skip the check there too when it's turned off — matching the main scale-up path.
See PR (to follow) for the implementation.
Description
The main scale-up path has a config flag,
ENABLE_JOB_QUEUED_CHECK, that skips an API call to GitHub that checks whether a job is still queued before creating a runner for it. The retry lambda does not check this flag — it always makes that call, on every retry attempt.Impact
Turning off
ENABLE_JOB_QUEUED_CHECKdoes not fully turn off the check: the retry path keeps calling GitHub's API for every retry, using up rate-limit budget the flag was meant to save. This makes the flag behave inconsistently depending on which lambda handles a given job.Proposed fix
Read the same
ENABLE_JOB_QUEUED_CHECKflag in the retry lambda, and skip the check there too when it's turned off — matching the main scale-up path.See PR (to follow) for the implementation.