Investigation of double distribution occurance

[cielf]

Summary

Investigation of the issue covered in the support log at Nov 12, 2024, 11:17.

Please check into the logs and see if there are any further clues there.

We had thought this was a double-click on the distribution confirmation, but there has been a fix put in to cover off that situation.

Things to consider

No response

Criteria for Completion

Investigate and report back to the Sunday planning meeting.

[dorner]

I am definitely seeing multiple requests coming in at the exact same time.

This seems to be because the request timed out. I'm seeing a request taking over 48 seconds and nginx returns an error 499. I'll bet that this causes the button to re-enable and the user clicks again at that point.

We need to fix our slow queries.

EDIT: Validated that when this happens the user actually gets an error screen. They can then click back and go back to the request page and do everything again since at that point the request hasn't been fulfilled yet. That seems to be what happened.

[dorner]

I'm wondering if this has something to do with vacuuming - maybe it happens to be going on at that exact moment.

[cielf]

Noting that during the planning meeting we determined that vacuuming wasn't happening at the time of the issue.

[github-actions]

This issue is marked as stale due to no activity within 30 days. If no further activity is detected within 7 days, it will be unassigned.

[github-actions]

Automatically unassigned after 7 days of inactivity.

[github-actions]

This issue is marked as stale due to no activity within 30 days. If no further activity is detected within 7 days, it will be unassigned.

[github-actions]

Automatically unassigned after 7 days of inactivity.

[dorner]

Going to close this for now since the weirdo query times are still showing up on Newrelic but don't seem to affect app times. If we see this problem again, we can open a new issue for it.