feat(stats): add HyStart++ stats#3432
Merged
Merged
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #3432 +/- ##
==========================================
- Coverage 94.30% 94.17% -0.13%
==========================================
Files 127 131 +4
Lines 38708 39041 +333
Branches 38708 39041 +333
==========================================
+ Hits 36502 36768 +266
- Misses 1365 1424 +59
- Partials 841 849 +8
Flags with carried forward coverage won't be shown. Click here to find out more.
|
Merging this PR will not alter performance
Comparing 
Footnotes
|
- `CongestionEvent` if loss or ecn - `Heuristic` if algorithm (e.g. HyStart++)
omansfeld
force-pushed
the
hystart_metrics
branch
from
f401181 to
1b8ad69
Compare
omansfeld
requested review from
KershawChang,
larseggert,
martinthomson and
mxinden
as code owners
Contributor
There was a problem hiding this comment.
Pull request overview
Adds additional congestion-control statistics to better observe HyStart++ behavior and slow-start exits, integrating them into the transport stats surface and updating CC/tests accordingly.
Changes:
- Introduces
SlowStartExitReasonand recordsslow_start_exit_reasonwhen leaving slow start via congestion events vs heuristic exit. - Adds HyStart++ CSS counters (
css_entries,css_rounds_finished) and increments them inside the HyStart++ implementation. - Updates debug formatting and HyStart++/classic CC tests to cover the new stats.
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
neqo-transport/src/stats.rs |
Adds the new exit reason enum + new congestion control stats fields; updates Stats debug output. |
neqo-transport/src/cc/hystart.rs |
Increments CSS entry/round counters during HyStart++ processing. |
neqo-transport/src/cc/classic_slow_start.rs |
Updates the slow start trait impl signature to accept stats mutably. |
neqo-transport/src/cc/classic_cc.rs |
Records slow_start_exit_reason for heuristic exits and congestion events; updates tests accordingly. |
neqo-transport/src/cc/tests/hystart.rs |
Extends HyStart++ tests to validate the new stats. |
You can also share your feedback on Copilot code review. Take the survey.
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> Signed-off-by: Oskar Mansfeld <dev@omansfeld.net>
Contributor
Client/server transfer resultsPerformance differences relative to 6a953c2. Transfer of 33554432 bytes over loopback, min. 100 runs. All unit-less numbers are in milliseconds.
Table above only shows statistically significant changes. See all results below. All resultsTransfer of 33554432 bytes over loopback, min. 100 runs. All unit-less numbers are in milliseconds.
Download data for |
Signed-off-by: Oskar Mansfeld <git@omansfeld.net>
mxinden
approved these changes
Mar 16, 2026
Contributor
Benchmark resultsSignificant performance differences relative to bf77765. transfer/walltime/pacing-false/same-seed: 💔 Performance has regressed by +3.2665%. time: [80.279 ms 80.340 ms 80.406 ms]
change: [+3.0475% +3.2665% +3.4320] (p = 0.00 < 0.05)
Performance has regressed.
Found 4 outliers among 100 measurements (4.00%)
2 (2.00%) high mild
2 (2.00%) high severeAll resultstransfer/1-conn/1-100mb-resp (aka. Download)/mtu-1504: No change in performance detected. time: [202.18 ms 202.55 ms 202.94 ms]
thrpt: [492.76 MiB/s 493.70 MiB/s 494.61 MiB/s]
change:
time: [-0.6060% -0.3014% +0.0053] (p = 0.05 > 0.05)
thrpt: [-0.0053% +0.3023% +0.6097]
No change in performance detected.
Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) high severetransfer/1-conn/10_000-parallel-1b-resp (aka. RPS)/mtu-1504: No change in performance detected. time: [284.40 ms 286.19 ms 288.01 ms]
thrpt: [34.721 Kelem/s 34.942 Kelem/s 35.162 Kelem/s]
change:
time: [-0.5411% +0.4177% +1.3542] (p = 0.39 > 0.05)
thrpt: [-1.3361% -0.4160% +0.5440]
No change in performance detected.
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mildtransfer/1-conn/1-1b-resp (aka. HPS)/mtu-1504: No change in performance detected. time: [38.552 ms 38.725 ms 38.919 ms]
thrpt: [25.694 B/s 25.823 B/s 25.939 B/s]
change:
time: [-0.4553% +0.1236% +0.7920] (p = 0.69 > 0.05)
thrpt: [-0.7858% -0.1234% +0.4573]
No change in performance detected.
Found 11 outliers among 100 measurements (11.00%)
2 (2.00%) low mild
3 (3.00%) high mild
6 (6.00%) high severetransfer/1-conn/1-100mb-req (aka. Upload)/mtu-1504: Change within noise threshold. time: [203.69 ms 204.16 ms 204.75 ms]
thrpt: [488.40 MiB/s 489.82 MiB/s 490.94 MiB/s]
change:
time: [-1.1756% -0.8715% -0.5155] (p = 0.00 < 0.05)
thrpt: [+0.5181% +0.8792% +1.1896]
Change within noise threshold.
Found 2 outliers among 100 measurements (2.00%)
1 (1.00%) high mild
1 (1.00%) high severestreams/walltime/1-streams/each-1000-bytes: Change within noise threshold. time: [590.75 µs 593.37 µs 596.28 µs]
change: [+0.4261% +1.1468% +1.8402] (p = 0.00 < 0.05)
Change within noise threshold.
Found 16 outliers among 100 measurements (16.00%)
2 (2.00%) low mild
14 (14.00%) high severestreams/walltime/1000-streams/each-1-bytes: Change within noise threshold. time: [12.430 ms 12.451 ms 12.473 ms]
change: [+0.3085% +0.7733% +1.1301] (p = 0.00 < 0.05)
Change within noise threshold.
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mildstreams/walltime/1000-streams/each-1000-bytes: No change in performance detected. time: [44.712 ms 44.807 ms 44.957 ms]
change: [-0.3287% -0.0829% +0.2467] (p = 0.66 > 0.05)
No change in performance detected.
Found 4 outliers among 100 measurements (4.00%)
1 (1.00%) low mild
2 (2.00%) high mild
1 (1.00%) high severetransfer/walltime/pacing-false/varying-seeds: Change within noise threshold. time: [78.912 ms 79.019 ms 79.172 ms]
change: [+1.3339% +1.4967% +1.7261] (p = 0.00 < 0.05)
Change within noise threshold.
Found 6 outliers among 100 measurements (6.00%)
1 (1.00%) low mild
3 (3.00%) high mild
2 (2.00%) high severetransfer/walltime/pacing-true/varying-seeds: Change within noise threshold. time: [80.308 ms 80.406 ms 80.543 ms]
change: [+0.9533% +1.1907% +1.4058] (p = 0.00 < 0.05)
Change within noise threshold.
Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high severetransfer/walltime/pacing-false/same-seed: 💔 Performance has regressed by +3.2665%. time: [80.279 ms 80.340 ms 80.406 ms]
change: [+3.0475% +3.2665% +3.4320] (p = 0.00 < 0.05)
Performance has regressed.
Found 4 outliers among 100 measurements (4.00%)
2 (2.00%) high mild
2 (2.00%) high severetransfer/walltime/pacing-true/same-seed: Change within noise threshold. time: [80.345 ms 80.450 ms 80.605 ms]
change: [+0.9532% +1.1326% +1.3631] (p = 0.00 < 0.05)
Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
2 (2.00%) high mild
2 (2.00%) high severeDownload data for |
Contributor
Failed Interop TestsQUIC Interop Runner, client vs. server, differences relative to
All resultsSucceeded Interop TestsQUIC Interop Runner, client vs. server neqo-pr as client
neqo-pr as server
Unsupported Interop TestsQUIC Interop Runner, client vs. server neqo-pr as client
neqo-pr as server
|
This was referenced Mar 16, 2026
Merged
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Depends on #3278. I will rebase once that lands on main. Only commits belonging to this PR are the following:add
slow_start_exit_reasonstatCongestionEventif loss or ecnHeuristicif algorithm (e.g. HyStart++)add
css_rounds_finishedstatadd
css_entriesstat to count how often css was enteredThe two CSS stats can be used to judge how often HyStart++ spuriously enters CSS and thus spends (unnecessary) time in throttled growth. In a perfect world we'd have 1 css entry and 5 rounds finished. If values are wildly increased that might point to an inefficiency.
The exit reason stat can be used to judge how well HyStart++ (or another heuristic) performs at detecting an exit point before packet loss occurs. Can also be used as a filter when recording stats on the firefox side as glean probes, e.g. to differentiate
slow_start_exit_accuracy(introduced in this bug) for heuristic and congestion event based exits.