diff options
| author | Taras Madan <tarasmadan@google.com> | 2024-07-26 12:52:50 +0200 |
|---|---|---|
| committer | Taras Madan <tarasmadan@google.com> | 2024-07-26 12:29:10 +0000 |
| commit | 7e805b6a42655e3b327f820cd56f88b99ab4db2a (patch) | |
| tree | 35539ca97aec3f7d641b7c0b60f1307677dbff30 /executor/common_ext_example.h | |
| parent | 389c3e9cea0e1f2b34d475e858942ae9c6f12e05 (diff) | |
dashboard/app/coverage_batch.go: don't retry even on spot
When running on Spot VMs, our batch jobs can get preempted and restarted.
For some reason, GCP sometimes starts multiple instances of the same job,
which results in duplicate results in the DB.
Prevent this by dropping MaxRetryCount. Even if a job fails for a specific day,
the data for that day will be processed on the following day, so it does not get lost.
If needed, we can manually trigger recalc using /cron/batch_coverage.
Diffstat (limited to 'executor/common_ext_example.h')
0 files changed, 0 insertions, 0 deletions