[PLPeeters]

What do these changes do?

They replace per-release copying and shuffling of waiting host keys with rotation through an OrderedDict. Draining one waiter from each of N eligible host queues now takes O(N) scheduling work instead of O(N²).

Local benchmarks showed a 33x speedup for 100 hosts and a 535x speedup for 1600 hosts (yes, you read that correctly).

Are there changes in behavior for the user?

Cross-host scheduling changes from randomized selection to round-robin rotation. Connection limits, FIFO order within each host, and cancellation handling are preserved.

Clients queuing requests across many distinct hosts should spend less time scheduling waiting connections.

Is it a substantial burden for the maintainers to support this?

No new dependencies or public APIs are introduced. A single ordered mapping holds the host queues and their scheduling order, avoiding duplicate bookkeeping.

The main maintenance concern is preserving queue cleanup and rotation during cancellation and retries; regression tests cover these paths.

Related issue number

N/A

Checklist

• I think the code is well written
• Unit tests for the changes exist
• Documentation reflects the changes
• If you provide code modifications, please add yourself to
  CONTRIBUTORS.txt
• Add a new news fragment into the CHANGES/ folder

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.12%. Comparing base (fe09949) to head (c1720ba).
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #14379      +/-   ##
==========================================
- Coverage   99.12%   99.12%   -0.01%     
==========================================
  Files         135      136       +1     
  Lines       53661    53713      +52     
  Branches     2737     2741       +4     
==========================================
+ Hits        53191    53242      +51     
  Misses        356      356              
- Partials      114      115       +1     

Flag	Coverage Δ
Autobahn	21.66% <20.00%> (-0.01%)	⬇️
CI-GHA	99.05% <100.00%> (-0.01%)	⬇️
OS-Linux	98.73% <100.00%> (-0.01%)	⬇️
OS-Windows	97.48% <100.00%> (-0.01%)	⬇️
OS-macOS	98.34% <100.00%> (-0.01%)	⬇️
Py-3.11	98.29% <100.00%> (-0.01%)	⬇️
Py-3.12	98.37% <100.00%> (-0.01%)	⬇️
Py-3.13	98.36% <100.00%> (-0.01%)	⬇️
Py-3.14	98.42% <100.00%> (-0.01%)	⬇️
Py-3.15	98.43% <100.00%> (+<0.01%)	⬆️
Py-3.15t	98.27% <100.00%> (-0.02%)	⬇️
Py-pypy-3.12	96.95% <100.00%> (-0.02%)	⬇️
VM-macos-latest	98.34% <100.00%> (-0.01%)	⬇️
VM-ubuntu-latest	98.73% <100.00%> (-0.01%)	⬇️
VM-windows-latest	97.48% <100.00%> (-0.01%)	⬇️
cython-coverage	83.93% <9.09%> (-0.02%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[codspeed]

Merging this PR will not alter performance

1 warning

⚠️ 50 benchmarks are dominated by system calls

System calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.

Learn more about system calls

See affected benchmarks

✅ 102 untouched benchmarks
🆕 3 new benchmarks
⏩ 83 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
🆕	test_release_waiters_across_distinct_hosts[1600]	N/A	32.2 ms	N/A
🆕	test_release_waiters_across_distinct_hosts[400]	N/A	8.1 ms	N/A
🆕	test_release_waiters_across_distinct_hosts[800]	N/A	16.2 ms	N/A

---

_{Comparing PLPeeters:bugfix/quadtratic-connector-wait (c1720ba) with master (fe09949)}

Footnotes

1. 83 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[PLPeeters]

@codspeed Can you run the new bench tests against master so we have a comparison?

[codspeed]

@codspeed Can you run the new bench tests against master so we have a comparison?

To let the performance wizard handle your request, please sign in to CodSpeed at codspeed.io so we can link your GitHub account, then comment again.

[PLPeeters]

@codspeed Can you run the new bench tests against master so we have a comparison?

[codspeed]

@codspeed Can you run the new bench tests against master so we have a comparison?

Only members of the aio-libs organization on CodSpeed can ask the performance wizard for help here. Once you've joined it on CodSpeed, comment again.

[PLPeeters]

Saw that one coming but had to try haha