Repository navigation
roadrunner: update to v0.13.1 - #1673
Open
williamthome wants to merge 4 commits into
Open
williamthome wants to merge 4 commits into
williamthome wants to merge 4 commits into
Conversation
williamthome
requested review from
Kaliumhexacyanoferrat and
MDA2AV
as code owners
October 28, 2026 18:29
Contributor
Author
|
/benchmark-multiple -f roadrunner,roadrunner-gateway |
Contributor
|
👋 Benchmark request received. A collaborator will review and approve the run. |
Contributor
Benchmark ResultsFrameworks: 2 | Test: ✅
|
| Test | Conn | RPS | Rate | CPU/req | p99 | p99.9 | CPU | Mem | Δ RPS | Δ Mem |
|---|---|---|---|---|---|---|---|---|---|---|
| baseline | 4096 | 1,148,070 | — | — | — | — | 6286.4% | 331MiB | -1.6% | -2.9% |
| pipelined | 4096 | 5,311,003 | — | — | — | — | 6165.9% | 256MiB | -13.4% | -3.4% |
| limited-conn | 4096 | 459,258 | — | — | — | — | 4191.9% | 231MiB | +1.9% | -0.9% |
| json-comp | 4096 | 111,370 | — | — | — | — | 6264.0% | 659MiB | +6.0% | +6.1% |
| json-comp | 16384 | 113,369 | — | — | — | — | 6448.8% | 671MiB | +3.1% | +12.2% |
| json-tls | 4096 | 167,736 | — | — | — | — | 6423.6% | 898MiB | +1.6% | -0.3% |
| 8gbit | 512 | 48,341 | 0.9668 |
278.46us | 1297.0us | 6114.0us | 1273.8% | 462MiB | -0.2% | -4.9% |
| static-tls | 1024 | 83,237 | — | — | — | — | 6165.0% | 699MiB | -5.1% | -0.1% |
| async-db | 1024 | 125,861 | — | — | — | — | 6432.3% | 716MiB | -5.9% | -5.2% |
| fortunes | 1024 | 69,440 | — | — | — | — | 6250.4% | 1.2GiB | -1.2% | +29.5% |
| baseline-h2 | 256 | 602,386 | — | — | — | — | 6644.6% | 354MiB | +17.7% | +5.7% |
| baseline-h2 | 1024 | 603,819 | — | — | — | — | 6493.3% | 480MiB | +15.8% | +5.5% |
| static-h2 | 256 | 78,203 | — | — | — | — | 6050.3% | 488MiB | +0.5% | -7.9% |
| static-h2 | 1024 | 74,228 | — | — | — | — | 6269.2% | 999MiB | -1.5% | +0.6% |
| baseline-h2c | 256 | 966,093 | — | — | — | — | 6620.1% | 273MiB | +8.0% | +23.0% |
| baseline-h2c | 1024 | 942,458 | — | — | — | — | 6523.9% | 362MiB | +8.5% | +42.5% |
| baseline-h2c | 4096 | 873,164 | — | — | — | — | 6546.3% | 585MiB | +12.4% | +62.5% |
| json-h2c | 1024 | 113,317 | — | — | — | — | 3738.2% | 932MiB | +9.6% | +30.5% |
| json-h2c | 4096 | 113,958 | — | — | — | — | 4080.1% | 1.5GiB | -11.4% | +15.4% |
| baseline-h3 | 64 | 181,454 | — | — | — | — | 3754.4% | 376MiB | -1.2% | +4.2% |
| static-h3 | 64 | 47,376 | — | — | — | — | 4219.0% | 1.4GiB | +13.9% | +7.7% |
| echo-ws | 512 | 466,495 | — | — | — | — | 4763.7% | 210MiB | +24.9% | +3.4% |
| echo-ws | 4096 | 1,256,222 | — | — | — | — | 5613.5% | 387MiB | +23.6% | -0.3% |
| echo-ws | 16384 | 321,391 | — | — | — | — | 4328.5% | 1015MiB | -79.8% | +12.8% |
| echo-ws-pipeline | 512 | 17,937,465 | — | — | — | — | 6553.1% | 191MiB | +41.8% | -5.9% |
| echo-ws-pipeline | 4096 | 16,479,190 | — | — | — | — | 6343.3% | 308MiB | +33.3% | -1.9% |
| echo-ws-pipeline | 16384 | 16,572,620 | — | — | — | — | 6664.9% | 329MiB | +35.0% | -2.7% |
| echo-ws-limited | 512 | 435,361 | — | — | — | — | 4504.8% | 236MiB | +45.0% | -17.5% |
| echo-ws-limited | 4096 | 439,701 | — | — | — | — | 4301.7% | 232MiB | +54.5% | -55.9% |
| latency-1m | 1024 | 920,912 | 0.9209 |
64.12us | 1837568.0us | 2427904.0us | 6123.7% | 361MiB | +2.5% | +4.9% |
| latency-10k | 1024 | 9,979 | 0.9979 | 128.67us | 147.0us | 653.0us | 124.6% | 216MiB | ~0% | +5.4% |
| latency-500k-8cpu | 1024 | 176,100 | 0.3522 |
45.33us | 12955648.0us | 13217792.0us | 771.9% | 143MiB | +0.3% | +0.7% |
| async | 32000 | 1,142,223 | — | — | — | — | 6291.6% | 875MiB | -2.3% | +7.6% |
Composite score, main against this run, in the board's default view:
| Family | main | this run | Δ |
|---|---|---|---|
| HTTP/1.1 | 2443.0 (#54 of 102) | 2466.7 (#54 of 102) | +23.7 |
| HTTP/2 | 260.5 (#30 of 35) | 274.6 (#30 of 35) | +14.1 |
| HTTP/3 | 180.1 (#17 of 19) | 199.6 (#17 of 19) | +19.5 |
| Gateway | 2340.9 (#3 of 9) | 2337.7 (#3 of 9) | -3.2 |
| WebSocket | 597.9 (#26 of 35) | 664.6 (#25 of 35) | +66.8 |
The results on main were measured in another session, so a few points up or down can be the machine and not the change.
✅ roadrunner-gateway
| Test | Conn | RPS | CPU | Mem | Δ RPS | Δ Mem |
|---|---|---|---|---|---|---|
| gateway-64 | 512 | 155,994 | 5929.4% | 2.9GiB | +0.5% | ~0% |
| gateway-64 | 1024 | 151,295 | 5958.2% | 4.4GiB | +0.3% | +2.3% |
| gateway-h3 | 64 | 81,666 | 5286.6% | 770MiB | +0.2% | -5.5% |
| gateway-h3 | 256 | 58,978 | 4517.6% | 1.6GiB | +0.7% | -5.9% |
| production-stack | 256 | 236,237 | 5753.8% | 1.4GiB | +2.1% | ~0% |
| production-stack | 1024 | 185,500 | 5418.2% | 2.1GiB | -4.6% | -8.7% |
Contributor
Author
|
/benchmark -f roadrunner -t echo-ws |
Contributor
|
👋 Benchmark request received. A collaborator will review and approve the run. |
Contributor
Benchmark ResultsFramework:
Composite score, main against this run, in the board's default view:
The results on main were measured in another session, so a few points up or down can be the machine and not the change. Full log |
Contributor
Author
|
/benchmark -f roadrunner -t echo-ws |
Contributor
|
👋 Benchmark request received. A collaborator will review and approve the run. |
Contributor
Benchmark ResultsFramework:
Composite score, main against this run, in the board's default view:
The results on main were measured in another session, so a few points up or down can be the machine and not the change. Full log |
Contributor
Author
|
/benchmark-multiple -f roadrunner,roadrunner-gateway |
Contributor
|
👋 Benchmark request received. A collaborator will review and approve the run. |
Contributor
Benchmark ResultsFrameworks: 2 | Test: ✅
|
| Test | Conn | RPS | Rate | CPU/req | p99 | p99.9 | CPU | Mem | Δ RPS | Δ Mem |
|---|---|---|---|---|---|---|---|---|---|---|
| baseline | 4096 | 1,178,019 | — | — | — | — | 6399.5% | 360MiB | +1.0% | +5.6% |
| pipelined | 4096 | 5,427,942 | — | — | — | — | 6200.7% | 268MiB | -11.5% | +1.1% |
| limited-conn | 4096 | 517,705 | — | — | — | — | 3472.3% | 224MiB | +14.8% | -3.9% |
| json-comp | 4096 | 103,671 | — | — | — | — | 6445.4% | 632MiB | -1.3% | +1.8% |
| json-comp | 16384 | 95,801 | — | — | — | — | 6455.2% | 673MiB | -12.8% | +12.5% |
| json-tls | 4096 | 184,269 | — | — | — | — | 6367.7% | 980MiB | +11.6% | +8.8% |
| 8gbit | 512 | 48,449 | 0.9690 |
272.35us | 1114.0us | 4558.0us | 1241.3% | 476MiB | ~0% | -2.1% |
| static-tls | 1024 | 86,715 | — | — | — | — | 6050.4% | 658MiB | -1.2% | -6.0% |
| async-db | 1024 | 132,058 | — | — | — | — | 6418.6% | 724MiB | -1.3% | -4.1% |
| fortunes | 1024 | 71,074 | — | — | — | — | 6409.1% | 1.1GiB | +1.1% | +18.7% |
| baseline-h2 | 256 | 594,451 | — | — | — | — | 6650.9% | 341MiB | +16.1% | +1.8% |
| baseline-h2 | 1024 | 606,937 | — | — | — | — | 6479.0% | 464MiB | +16.4% | +2.0% |
| static-h2 | 256 | 76,529 | — | — | — | — | 6241.7% | 501MiB | -1.6% | -5.5% |
| static-h2 | 1024 | 78,214 | — | — | — | — | 6004.0% | 990MiB | +3.8% | -0.3% |
| baseline-h2c | 256 | 945,225 | — | — | — | — | 6561.3% | 251MiB | +5.7% | +13.1% |
| baseline-h2c | 1024 | 948,059 | — | — | — | — | 6521.9% | 361MiB | +9.1% | +42.1% |
| baseline-h2c | 4096 | 856,788 | — | — | — | — | 6372.0% | 577MiB | +10.3% | +60.3% |
| json-h2c | 1024 | 115,659 | — | — | — | — | 3883.2% | 883MiB | +11.9% | +23.7% |
| json-h2c | 4096 | 114,631 | — | — | — | — | 4089.4% | 1.5GiB | -10.9% | +15.4% |
| baseline-h3 | 64 | 182,100 | — | — | — | — | 3852.6% | 353MiB | -0.8% | -2.2% |
| static-h3 | 64 | 43,903 | — | — | — | — | 3576.1% | 1.3GiB | +5.6% | ~0% |
| echo-ws | 512 | 1,982,760 | — | — | — | — | 6623.8% | 217MiB | +430.9% | +6.9% |
| echo-ws | 4096 | 1,944,216 | — | — | — | — | 6519.4% | 255MiB | +91.2% | -34.3% |
| echo-ws | 16384 | 1,792,260 | — | — | — | — | 6294.3% | 356MiB | +12.7% | -60.4% |
| echo-ws-pipeline | 512 | 17,475,308 | — | — | — | — | 6534.3% | 236MiB | +38.2% | +16.3% |
| echo-ws-pipeline | 4096 | 16,904,531 | — | — | — | — | 6523.7% | 250MiB | +36.8% | -20.4% |
| echo-ws-pipeline | 16384 | 16,453,388 | — | — | — | — | 6646.6% | 333MiB | +34.0% | -1.5% |
| echo-ws-limited | 512 | 533,400 | — | — | — | — | 3137.1% | 231MiB | +77.7% | -19.2% |
| echo-ws-limited | 4096 | 527,074 | — | — | — | — | 3277.0% | 245MiB | +85.2% | -53.4% |
| latency-1m | 1024 | 997,093 | 0.9971 | 62.65us | 30024.0us | 65568.0us | 6497.4% | 454MiB | +11.0% | +32.0% |
| latency-10k | 1024 | 9,982 | 0.9982 | 125.80us | 135.0us | 840.0us | 121.0% | 204MiB | ~0% | -0.5% |
| latency-500k-8cpu | 1024 | 174,265 | 0.3485 |
45.78us | 12980224.0us | 13275136.0us | 769.5% | 138MiB | -0.7% | -2.8% |
| async | 32000 | 1,146,710 | — | — | — | — | 6350.8% | 882MiB | -1.9% | +8.5% |
Composite score, main against this run, in the board's default view:
| Family | main | this run | Δ |
|---|---|---|---|
| HTTP/1.1 | 2443.0 (#54 of 102) | 2580.8 (#54 of 102) | +137.8 |
| HTTP/2 | 260.5 (#30 of 35) | 275.3 (#30 of 35) | +14.8 |
| HTTP/3 | 180.1 (#17 of 19) | 187.7 (#17 of 19) | +7.7 |
| Gateway | 2340.9 (#3 of 9) | 2341.2 (#3 of 9) | +0.3 |
| WebSocket | 597.9 (#26 of 35) | 1008.8 (#21 of 35) | +411.0 |
The results on main were measured in another session, so a few points up or down can be the machine and not the change.
✅ roadrunner-gateway
| Test | Conn | RPS | CPU | Mem | Δ RPS | Δ Mem |
|---|---|---|---|---|---|---|
| gateway-64 | 512 | 155,800 | 5945.3% | 2.9GiB | +0.4% | ~0% |
| gateway-64 | 1024 | 151,071 | 6014.5% | 4.2GiB | +0.1% | -2.3% |
| gateway-h3 | 64 | 81,717 | 5362.7% | 769MiB | +0.3% | -5.6% |
| gateway-h3 | 256 | 58,670 | 4228.3% | 1.2GiB | +0.2% | -29.4% |
| production-stack | 256 | 227,800 | 5451.1% | 1.4GiB | -1.5% | ~0% |
| production-stack | 1024 | 196,642 | 5495.6% | 2.1GiB | +1.1% | -8.7% |
Contributor
Author
|
/benchmark-multiple -f roadrunner,roadrunner-gateway --save |
Contributor
|
👋 Benchmark request received. A collaborator will review and approve the run. |
Contributor
|
|
williamthome
changed the title
williamthome
force-pushed
the
entries/roadrunner-0.13.0
branch
from
October 29, 2026 12:13
d54b2db to
cba8308
Compare
Contributor
Author
|
/benchmark-multiple -f roadrunner,roadrunner-gateway --save |
Contributor
|
👋 Benchmark request received. A collaborator will review and approve the run. |
Contributor
Benchmark ResultsFrameworks: 2 | Test: ✅
|
| Test | Conn | RPS | Rate | CPU/req | p99 | p99.9 | CPU | Mem | Δ RPS | Δ Mem |
|---|---|---|---|---|---|---|---|---|---|---|
| baseline | 4096 | 1,189,342 | — | — | — | — | 6435.6% | 360MiB | +2.0% | +5.6% |
| pipelined | 4096 | 5,821,942 | — | — | — | — | 6288.5% | 242MiB | -5.1% | -8.7% |
| limited-conn | 4096 | 518,855 | — | — | — | — | 3586.0% | 232MiB | +15.1% | -0.4% |
| json-comp | 4096 | 111,101 | — | — | — | — | 6250.0% | 641MiB | +5.8% | +3.2% |
| json-comp | 16384 | 104,738 | — | — | — | — | 6105.0% | 693MiB | -4.7% | +15.9% |
| json-tls | 4096 | 165,677 | — | — | — | — | 6434.0% | 863MiB | +0.4% | -4.2% |
| 8gbit | 512 | 48,479 | 0.9696 |
280.06us | 711.0us | 7862.0us | 1286.8% | 475MiB | +0.1% | -2.3% |
| static-tls | 1024 | 83,444 | — | — | — | — | 5770.8% | 711MiB | -4.9% | +1.6% |
| async-db | 1024 | 122,727 | — | — | — | — | 6449.3% | 724MiB | -8.3% | -4.1% |
| fortunes | 1024 | 70,530 | — | — | — | — | 6470.5% | 915MiB | +0.3% | -3.6% |
| baseline-h2 | 256 | 605,917 | — | — | — | — | 6565.0% | 360MiB | +18.4% | +7.5% |
| baseline-h2 | 1024 | 607,831 | — | — | — | — | 6505.6% | 479MiB | +16.6% | +5.3% |
| static-h2 | 256 | 79,399 | — | — | — | — | 6114.4% | 528MiB | +2.0% | -0.4% |
| static-h2 | 1024 | 75,454 | — | — | — | — | 6188.5% | 1.0GiB | +0.1% | +3.1% |
| baseline-h2c | 256 | 962,714 | — | — | — | — | 6586.5% | 263MiB | +7.6% | +18.5% |
| baseline-h2c | 1024 | 948,969 | — | — | — | — | 6553.6% | 371MiB | +9.2% | +46.1% |
| baseline-h2c | 4096 | 869,082 | — | — | — | — | 6553.1% | 631MiB | +11.8% | +75.3% |
| json-h2c | 1024 | 110,208 | — | — | — | — | 3856.4% | 895MiB | +6.6% | +25.4% |
| json-h2c | 4096 | 103,104 | — | — | — | — | 4219.4% | 1.3GiB | -19.8% | ~0% |
| baseline-h3 | 64 | 184,160 | — | — | — | — | 4100.3% | 322MiB | +0.3% | -10.8% |
| static-h3 | 64 | 37,532 | — | — | — | — | 3267.3% | 1.3GiB | -9.8% | ~0% |
| echo-ws | 512 | 1,969,616 | — | — | — | — | 6600.8% | 195MiB | +427.4% | -3.9% |
| echo-ws | 4096 | 1,960,375 | — | — | — | — | 6726.2% | 266MiB | +92.8% | -31.4% |
| echo-ws | 16384 | 1,863,889 | — | — | — | — | 6644.8% | 342MiB | +17.2% | -62.0% |
| echo-ws-pipeline | 512 | 17,792,009 | — | — | — | — | 6641.0% | 190MiB | +40.7% | -6.4% |
| echo-ws-pipeline | 4096 | 16,772,992 | — | — | — | — | 6426.1% | 304MiB | +35.7% | -3.2% |
| echo-ws-pipeline | 16384 | 16,408,758 | — | — | — | — | 6579.2% | 329MiB | +33.6% | -2.7% |
| echo-ws-limited | 512 | 537,168 | — | — | — | — | 3236.0% | 238MiB | +78.9% | -16.8% |
| echo-ws-limited | 4096 | 537,818 | — | — | — | — | 2986.5% | 270MiB | +89.0% | -48.7% |
| latency-1m | 1024 | 997,824 | 0.9978 | 62.58us | 23016.0us | 39888.0us | 6493.2% | 482MiB | +11.0% | +40.1% |
| latency-10k | 1024 | 9,982 | 0.9982 | 133.25us | 135.0us | 277.0us | 129.1% | 193MiB | ~0% | -5.9% |
| latency-500k-8cpu | 1024 | 175,249 | 0.3505 |
45.54us | 12980224.0us | 13234176.0us | 772.0% | 146MiB | -0.2% | +2.8% |
| async | 32000 | 1,143,294 | — | — | — | — | 6357.7% | 899MiB | -2.2% | +10.6% |
Composite score, main against this run, in the board's default view:
| Family | main | this run | Δ |
|---|---|---|---|
| HTTP/1.1 | 2443.0 (#54 of 102) | 2603.4 (#54 of 102) | +160.4 |
| HTTP/2 | 260.5 (#30 of 35) | 273.1 (#30 of 35) | +12.6 |
| HTTP/3 | 180.1 (#17 of 19) | 166.1 (#17 of 19) | -13.9 |
| Gateway | 2340.9 (#3 of 9) | 2345.8 (#3 of 9) | +4.9 |
| WebSocket | 597.9 (#26 of 35) | 1019.0 (#21 of 35) | +421.1 |
The results on main were measured in another session, so a few points up or down can be the machine and not the change.
✅ roadrunner-gateway
| Test | Conn | RPS | CPU | Mem | Δ RPS | Δ Mem |
|---|---|---|---|---|---|---|
| gateway-64 | 512 | 154,234 | 5983.7% | 2.9GiB | -0.6% | ~0% |
| gateway-64 | 1024 | 150,585 | 5873.6% | 4.6GiB | -0.2% | +7.0% |
| gateway-h3 | 64 | 81,244 | 5299.5% | 783MiB | -0.3% | -3.9% |
| gateway-h3 | 256 | 58,118 | 4536.9% | 1.7GiB | -0.8% | ~0% |
| production-stack | 256 | 236,088 | 5765.0% | 1.5GiB | +2.1% | +7.1% |
| production-stack | 1024 | 195,191 | 5482.6% | 2.1GiB | +0.3% | -8.7% |
Contributor
Author
|
@MDA2AV ready to merge, thanks! |
This branch has not been deployed
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.
Description
Updates roadrunner to v0.13.1, which reads HTTP/2 sockets with
{active, N}and HTTP/1.1 and WebSocket sockets with{active, N}on TLS (saving a call into the TLS connection process per read), handles complete WebSocket frames in one step, runs each WebSocket session in its connection process, and makes HTTP/3 sturdier under packet loss. Theroadrunner-gatewayentry builds fromframeworks/roadrunner, so it picks up the new version too.The entry also adds a
vm.argsthat runs one I/O poll thread, each with its own pollset, per eight schedulers (+IOPt 13 +IOPp 100), as roadrunner's resource limits guide recommends for many-core servers (docs/resource_limits.md). With the single default poll thread, a server with spare capacity on 64 schedulers was capped by that thread; the setting scales with the cpuset, so it gives 8 threads at 64 schedulers and keeps the default single thread below 16. The rest ofvm.argsis what relx generated before.PR Commands — comment on this PR to trigger (requires collaborator approval):
/benchmark -f <framework>/benchmark -f <framework> -t <test>/benchmark -f <framework> --save/benchmark -f <framework> -t <test> --save/benchmark -f <framework> --compare <other>/benchmark-multiple -f <fw1>,<fw2>,...-tand--savetoo; saved results land in a single commit/benchmark-multiple --save-fneeded: benchmark and save every framework the PR touches/benchmark-test -t <test><test>and save the resultsFor
/benchmark, always specify-f <framework>; the flags combine in any order. Results come back as a comment with a per-profile table of RPS, p99, CPU and memory — one table per framework on multi runs. A new benchmark comment while a run is in flight queues behind it (one deep) instead of cancelling it. For multi-framework PRs (dependency bumps, same-language refactors) prefer/benchmark-multiple, which runs everything in a single job and commits all saved results together, so no run overwrites another.--compareworks on single-framework runs only.What the deltas are measured against. By default, this framework's own results published on
main- answering "did this change help?". When you are tuning a variant or a successor entry,--comparere-bases them on another entry instead:The reply states which baseline it used, and profiles the other framework does not run show
n/arather than a delta.Run benchmarks locally
You can validate and benchmark your framework locally with the lite script — no CPU pinning, fixed connection counts, all load generators run in Docker.
Requirements: Docker Engine on Linux. Load generators (gcannon, h2load, h2load-h3, wrk) are built as self-contained Docker images on first run.