-
Notifications
You must be signed in to change notification settings - Fork 149
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix queue hang when collector exits without successfully benchmarking #1432
Comments
Regarding the input configuration (e.g.
I think that for fixing invalid input parameters, 1) would be preferable. |
IMO, (1) is nice-to-have validation, but we can't expect to duplicate all of that validation into the site -- that's not going to be maintainable. So I think pursuing (2) as the more general option is a better idea. |
Well site already depends on collector, so it wouldn't be that hard to expose a few functions that check the validity of CLI parameters (probably to avoid panicking in |
Sure, if that's easy we can do so. Some of the deeper validation - e.g., downloading the artifacts - feels a little harder to do well at that point, so I think we still want the more general feedback mechanism anyway. |
We have a general problem today that there's no general error feedback from the collector to the site -- only on a per-benchmark collection basis. That means that when there's some problem with the startup phase (e.g., no benchmarks match a filter, or SHA passed doesn't have artifacts), we end up hanging the queue until an operator manually intervenes and either removes or fixes the failing entry. (Currently done via manual database edit).
I think a good solution here would add some general error mechanism from the collector that could be posted back to the site. I think addressing the cases where the collector returns an Err is already 80% good enough -- we can work on avoiding panic! after that.
The text was updated successfully, but these errors were encountered: