Skip to content

Conversation

@wk9874
Copy link
Collaborator

@wk9874 wk9874 commented Feb 11, 2026

Hotfix: Metric size limits

Python Version(s) Tested: 3.11

Operating System(s): Ubuntu

📝 Summary

The total grid metric size limit was set to the same as the individual grid metric size limit (50k points, ~50KB). The total limit should be 10MB

This also exposed that the runs would go into lost state even though metrics were still being uploaded. Suspect this is due to the heartbeat thread being joined before the dispatcher thread was joined in _tidy_run or _error, meaning if there are lots of grids left to dispatch, the heartbeat would stop prematurely.

🔍 Diagnosis

Veeeery slow performance when loading large FDS simulations with lots of grid data. Run would become lost even while still uploading grid metrics.

🔄 Changes

Change the total metrics limit to be 1e6, and moved the joining of the heartbeat thread to below the joining of the dispatcher thread in _error and _tidy_run

✔️ Checklist

  • Unit and integration tests passing.
  • Pre-commit hooks passing.
  • Quality checks passing.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants