ir testsuite `stress_tests` seems to sometimes spuriously fail

The CI for an earlier version of #205 failed in a way that I could not reproduce locally. Testrun is [here](https://github.com/awslabs/AutoCorrode/actions/runs/24344158266/job/71080403913).

The relevant lines are:
```
Running stress tests (100 runs, 20 max concurrent, 10% rude disconnects)
↻ stress
✗ stress (381.6s)
1/100 runs failed: run 11: timed out
↻ rude_disconnect
✓ rude_disconnect (360.8s)

✗ 31/32 passed, 1 failed
Error: Process completed with exit code 1.
```

I am pretty sure that this failure is unrelated to the changes of #205. It feels like a rare racecondition, but I haven't investigated further.

The issue identifies the `stress_tests` suite and links to the GitHub Actions run; start by reviewing that run and its stress-test failure details. Determine why run 11 can time out and make the suite reliable, with CI no longer showing the reported spurious failure.

Written by the indexing model from the issue text.