[Bug]: Deleting a scanner query rule while the query is running generates a lot of tracebacks

[eviljeff]

Is there an existing issue for this?

• I have searched the existing issues

What happened?

https://mozilla.sentry.io/issues/7302612480/?project=6310819&query=is%3Aunresolved&referrer=issue-stream

A running scanner query resulted in 60 tracebacks in Sentry - which then triggered a critical error warning message to Slack - because the associated scanner query rule was deleted.

What did you expect to happen?

The query rule instance not existing was handled, so no exception was thrown. Or the rule was prevented from being deleted.

┆Issue is synchronized with this Jira Task

[diox]

Note: raising an error in this case was an explicit conscious decision at the time, born out of paranoia: we didn't want a scanner query rule execution to miss anything silently during normal operations.

We should look into:

• Properly cancelling the remaining tasks when a query rule is aborted or deleted (Implementing this FIXME)
  • Note that it will be too late for some tasks (they would already have been picked up by the workers), so this alone isn't enough, and might not be worth spending too much effort on.
• Silently handling the case where the rule doesn't exist