This doesn't look like it's an issue in this repo, but when forked I'm seeing regular action runs that stall for a full 6 hours and then timeout.
Not debilitating from the proposal perspective, but still a drain on resources.
See this example where a simple update to the README stalled out for 6h. Re-running failed jobs succeeded in <2m. Very possible that this is user error, so happy to be corrected if I missed an instruction somewhere.