fix: Handle errors from event stream callbacks

[tpoliaw]

If a callback raises an exception, it shouldn't prevent other callbacks
receiving the same event. Instead, catch any exceptions and re-raise
them after all callbacks have been called as a single ExceptionGroup.
This allows the task to be aborted if any of the callbacks fail but
still allows callbacks that require the "plan failed" events to run.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.00%. Comparing base (ae766ca) to head (7d6cd80).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1302      +/-   ##
==========================================
+ Coverage   94.99%   95.00%   +0.01%     
==========================================
  Files          42       42              
  Lines        2755     2763       +8     
==========================================
+ Hits         2617     2625       +8     
  Misses        138      138              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[tpoliaw]

The main symptom of this was that if a scan failed due to eg a stomp connection error, even if it reconnected, any subsequent scans would fail as the unsubscribe the tiled writer callback would never be called and tiled would return 409 conflict errors when two tiled writers tried to write the same events.

[abbiemery]

Demo'd in person

[abbiemery]

As it stands this will continue a scan running, even if rabbitmq is down. So we will loose data from the nexus file. Looking instead at a different way to get the errors to shut things down correctly.

Not desired behaviour.