Skip to content

Conversation

andrewlock
Copy link
Member

Summary of changes

Removes the Windows throughput tests

Reason for change

They've been broken for ~3 weeks, nobody's looking at them, they slow down the pipelines, and we're migrating to the benchmarking platform anyway

Implementation details

Delete the Windows job from the throughput stage - I've left all the other infra for it, as we will likely remove all of the crank stuff post- benchmarking platform transition

Test coverage

This is the test

Other details

Hopefully the compare stage works, this PR will show if it doesn't though

@andrewlock andrewlock added area:builds project files, build scripts, pipelines, versioning, releases, packages area:benchmarks Benchmarks, throughput tests, Crank, Bombardier, etc labels Mar 10, 2025
@andrewlock andrewlock requested a review from a team as a code owner March 10, 2025 14:43
@lucaspimentel lucaspimentel requested a review from a team March 10, 2025 16:23
@datadog-datadog-prod-us1
Copy link

datadog-datadog-prod-us1 bot commented Mar 10, 2025

Datadog Report

Branch report: andrew/ci/drop-windows-throughput-tests
Commit report: 723132c
Test service: dd-trace-dotnet

✅ 0 Failed, 245697 Passed, 2377 Skipped, 19h 40m 42.35s Total Time

@andrewlock
Copy link
Member Author

Execution-Time Benchmarks Report ⏱️

Execution-time results for samples comparing the following branches/commits:

Execution-time benchmarks measure the whole time it takes to execute a program. And are intended to measure the one-off costs. Cases where the execution time results for the PR are worse than latest master results are shown in red. The following thresholds were used for comparing the execution times:

  • Welch test with statistical test for significance of 5%
  • Only results indicating a difference greater than 5% and 5 ms are considered.

Note that these results are based on a single point-in-time result for each branch. For full results, see the dashboard.

Graphs show the p99 interval based on the mean and StdDev of the test run, as well as the mean value of the run (shown as a diamond below the graph).

gantt
    title Execution time (ms) FakeDbCommand (.NET Framework 4.6.2) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6752) - mean (69ms)  : 67, 72
     .   : milestone, 69,
    master - mean (69ms)  : 66, 72
     .   : milestone, 69,

    section CallTarget+Inlining+NGEN
    This PR (6752) - mean (1,005ms)  : 979, 1031
     .   : milestone, 1005,
    master - mean (1,003ms)  : 979, 1027
     .   : milestone, 1003,

Loading
gantt
    title Execution time (ms) FakeDbCommand (.NET Core 3.1) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6752) - mean (102ms)  : 100, 105
     .   : milestone, 102,
    master - mean (102ms)  : 100, 104
     .   : milestone, 102,

    section CallTarget+Inlining+NGEN
    This PR (6752) - mean (690ms)  : 672, 708
     .   : milestone, 690,
    master - mean (683ms)  : 666, 700
     .   : milestone, 683,

Loading
gantt
    title Execution time (ms) FakeDbCommand (.NET 6) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6752) - mean (89ms)  : 88, 91
     .   : milestone, 89,
    master - mean (89ms)  : 87, 91
     .   : milestone, 89,

    section CallTarget+Inlining+NGEN
    This PR (6752) - mean (645ms)  : 628, 661
     .   : milestone, 645,
    master - mean (640ms)  : 624, 657
     .   : milestone, 640,

Loading
gantt
    title Execution time (ms) HttpMessageHandler (.NET Framework 4.6.2) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6752) - mean (191ms)  : 186, 195
     .   : milestone, 191,
    master - mean (191ms)  : 187, 195
     .   : milestone, 191,

    section CallTarget+Inlining+NGEN
    This PR (6752) - mean (1,105ms)  : 1083, 1127
     .   : milestone, 1105,
    master - mean (1,108ms)  : 1084, 1131
     .   : milestone, 1108,

Loading
gantt
    title Execution time (ms) HttpMessageHandler (.NET Core 3.1) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6752) - mean (271ms)  : 267, 275
     .   : milestone, 271,
    master - mean (270ms)  : 267, 273
     .   : milestone, 270,

    section CallTarget+Inlining+NGEN
    This PR (6752) - mean (874ms)  : 845, 904
     .   : milestone, 874,
    master - mean (876ms)  : 850, 902
     .   : milestone, 876,

Loading
gantt
    title Execution time (ms) HttpMessageHandler (.NET 6) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6752) - mean (263ms)  : 259, 267
     .   : milestone, 263,
    master - mean (264ms)  : 257, 270
     .   : milestone, 264,

    section CallTarget+Inlining+NGEN
    This PR (6752) - mean (860ms)  : 828, 893
     .   : milestone, 860,
    master - mean (858ms)  : 823, 893
     .   : milestone, 858,

Loading

@andrewlock
Copy link
Member Author

Benchmarks Report for tracer 🐌

Benchmarks for #6752 compared to master:

  • 2 benchmarks are faster, with geometric mean 1.151
  • 1 benchmarks are slower, with geometric mean 1.157
  • All benchmarks have the same allocations

The following thresholds were used for comparing the benchmark speeds:

  • Mann–Whitney U test with statistical test for significance of 5%
  • Only results indicating a difference greater than 10% and 0.3 ns are considered.

Allocation changes below 0.5% are ignored.

Benchmark details

Benchmarks.Trace.ActivityBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master StartStopWithChild net6.0 8.1μs 47.4ns 452ns 0.0122 0.00406 0 5.61 KB
master StartStopWithChild netcoreapp3.1 10.9μs 62.2ns 494ns 0.0222 0.0111 0 5.8 KB
master StartStopWithChild net472 16.5μs 74.9ns 290ns 1.05 0.312 0.0959 6.22 KB
#6752 StartStopWithChild net6.0 8.51μs 47.8ns 328ns 0.016 0.008 0 5.61 KB
#6752 StartStopWithChild netcoreapp3.1 10.6μs 59.4ns 376ns 0.0155 0.00516 0 5.8 KB
#6752 StartStopWithChild net472 16.3μs 34.7ns 134ns 1.05 0.317 0.106 6.21 KB
Benchmarks.Trace.AgentWriterBenchmark - Faster 🎉 Same allocations ✔️

Faster 🎉 in #6752

Benchmark base/diff Base Median (ns) Diff Median (ns) Modality
Benchmarks.Trace.AgentWriterBenchmark.WriteAndFlushEnrichedTraces‑net6.0 1.153 552,783.13 479,611.57

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master WriteAndFlushEnrichedTraces net6.0 553μs 603ns 2.33μs 0 0 0 2.7 KB
master WriteAndFlushEnrichedTraces netcoreapp3.1 665μs 647ns 2.51μs 0 0 0 2.7 KB
master WriteAndFlushEnrichedTraces net472 865μs 291ns 1.01μs 0.428 0 0 3.3 KB
#6752 WriteAndFlushEnrichedTraces net6.0 480μs 461ns 1.72μs 0 0 0 2.7 KB
#6752 WriteAndFlushEnrichedTraces netcoreapp3.1 656μs 302ns 1.13μs 0 0 0 2.7 KB
#6752 WriteAndFlushEnrichedTraces net472 857μs 766ns 2.87μs 0.428 0 0 3.3 KB
Benchmarks.Trace.AspNetCoreBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master SendRequest net6.0 134μs 188ns 727ns 0.198 0 0 14.47 KB
master SendRequest netcoreapp3.1 154μs 126ns 470ns 0.153 0 0 17.27 KB
master SendRequest net472 0.000283ns 0.000181ns 0.000653ns 0 0 0 0 b
#6752 SendRequest net6.0 132μs 507ns 1.96μs 0.193 0 0 14.47 KB
#6752 SendRequest netcoreapp3.1 148μs 525ns 2.03μs 0.22 0 0 17.27 KB
#6752 SendRequest net472 0.00011ns 6.6E‑05ns 0.000256ns 0 0 0 0 b
Benchmarks.Trace.CIVisibilityProtocolWriterBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master WriteAndFlushEnrichedTraces net6.0 575μs 3.15μs 17.8μs 0.573 0 0 41.63 KB
master WriteAndFlushEnrichedTraces netcoreapp3.1 746μs 4.34μs 37.3μs 0.374 0 0 41.6 KB
master WriteAndFlushEnrichedTraces net472 848μs 3.27μs 12.7μs 8.7 2.77 0.396 53.33 KB
#6752 WriteAndFlushEnrichedTraces net6.0 556μs 3.1μs 19.4μs 0.546 0 0 41.67 KB
#6752 WriteAndFlushEnrichedTraces netcoreapp3.1 702μs 3.45μs 14.6μs 0.342 0 0 41.61 KB
#6752 WriteAndFlushEnrichedTraces net472 860μs 4.42μs 20.7μs 8.28 2.48 0.414 53.27 KB
Benchmarks.Trace.DbCommandBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master ExecuteNonQuery net6.0 1.4μs 1.93ns 7.46ns 0.0147 0 0 1.02 KB
master ExecuteNonQuery netcoreapp3.1 1.83μs 2.32ns 8.99ns 0.0136 0 0 1.02 KB
master ExecuteNonQuery net472 2.08μs 2.64ns 9.89ns 0.157 0.00105 0 987 B
#6752 ExecuteNonQuery net6.0 1.35μs 0.944ns 3.66ns 0.0142 0 0 1.02 KB
#6752 ExecuteNonQuery netcoreapp3.1 1.77μs 1.94ns 7.53ns 0.014 0 0 1.02 KB
#6752 ExecuteNonQuery net472 2.01μs 2.21ns 8.54ns 0.157 0.00101 0 987 B
Benchmarks.Trace.ElasticsearchBenchmark - Faster 🎉 Same allocations ✔️

Faster 🎉 in #6752

Benchmark base/diff Base Median (ns) Diff Median (ns) Modality
Benchmarks.Trace.ElasticsearchBenchmark.CallElasticsearchAsync‑net6.0 1.148 1,536.45 1,337.82

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master CallElasticsearch net6.0 1.27μs 0.443ns 1.66ns 0.014 0 0 976 B
master CallElasticsearch netcoreapp3.1 1.59μs 0.601ns 2.17ns 0.0131 0 0 976 B
master CallElasticsearch net472 2.42μs 1.87ns 7.25ns 0.158 0 0 995 B
master CallElasticsearchAsync net6.0 1.53μs 0.987ns 3.82ns 0.0131 0 0 952 B
master CallElasticsearchAsync netcoreapp3.1 1.73μs 0.598ns 2.31ns 0.0136 0 0 1.02 KB
master CallElasticsearchAsync net472 2.59μs 1.1ns 4.25ns 0.167 0 0 1.05 KB
#6752 CallElasticsearch net6.0 1.16μs 0.825ns 3.2ns 0.0134 0 0 976 B
#6752 CallElasticsearch netcoreapp3.1 1.58μs 0.618ns 2.14ns 0.0127 0 0 976 B
#6752 CallElasticsearch net472 2.48μs 1.14ns 4.13ns 0.158 0 0 995 B
#6752 CallElasticsearchAsync net6.0 1.34μs 0.605ns 2.34ns 0.0133 0 0 952 B
#6752 CallElasticsearchAsync netcoreapp3.1 1.68μs 0.863ns 3.34ns 0.0134 0 0 1.02 KB
#6752 CallElasticsearchAsync net472 2.63μs 1.3ns 5.02ns 0.166 0 0 1.05 KB
Benchmarks.Trace.GraphQLBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master ExecuteAsync net6.0 1.29μs 0.805ns 3.01ns 0.0135 0 0 952 B
master ExecuteAsync netcoreapp3.1 1.64μs 1ns 3.61ns 0.0123 0 0 952 B
master ExecuteAsync net472 1.88μs 0.424ns 1.64ns 0.145 0 0 915 B
#6752 ExecuteAsync net6.0 1.36μs 0.393ns 1.47ns 0.0132 0 0 952 B
#6752 ExecuteAsync netcoreapp3.1 1.58μs 0.57ns 2.13ns 0.0126 0 0 952 B
#6752 ExecuteAsync net472 1.86μs 0.407ns 1.47ns 0.145 0 0 915 B
Benchmarks.Trace.HttpClientBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master SendAsync net6.0 4.34μs 1.12ns 4.18ns 0.0327 0 0 2.31 KB
master SendAsync netcoreapp3.1 5.37μs 3.02ns 11.7ns 0.0374 0 0 2.85 KB
master SendAsync net472 7.53μs 2.15ns 8.05ns 0.494 0 0 3.12 KB
#6752 SendAsync net6.0 4.28μs 3.41ns 12.8ns 0.0319 0 0 2.31 KB
#6752 SendAsync netcoreapp3.1 5.5μs 2.89ns 11.2ns 0.0384 0 0 2.85 KB
#6752 SendAsync net472 7.46μs 1.55ns 6ns 0.492 0 0 3.12 KB
Benchmarks.Trace.ILoggerBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master EnrichedLog net6.0 1.45μs 0.884ns 3.31ns 0.0232 0 0 1.64 KB
master EnrichedLog netcoreapp3.1 2.2μs 1.48ns 5.34ns 0.0226 0 0 1.64 KB
master EnrichedLog net472 2.61μs 1.21ns 4.51ns 0.249 0 0 1.57 KB
#6752 EnrichedLog net6.0 1.56μs 1.17ns 4.36ns 0.0227 0 0 1.64 KB
#6752 EnrichedLog netcoreapp3.1 2.13μs 1.11ns 4.3ns 0.0223 0 0 1.64 KB
#6752 EnrichedLog net472 2.58μs 0.768ns 2.77ns 0.25 0 0 1.57 KB
Benchmarks.Trace.Log4netBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master EnrichedLog net6.0 112μs 204ns 789ns 0.0566 0 0 4.28 KB
master EnrichedLog netcoreapp3.1 117μs 208ns 780ns 0 0 0 4.28 KB
master EnrichedLog net472 148μs 206ns 770ns 0.665 0.222 0 4.46 KB
#6752 EnrichedLog net6.0 114μs 159ns 596ns 0 0 0 4.28 KB
#6752 EnrichedLog netcoreapp3.1 117μs 234ns 905ns 0.0587 0 0 4.28 KB
#6752 EnrichedLog net472 149μs 197ns 765ns 0.667 0.222 0 4.46 KB
Benchmarks.Trace.NLogBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master EnrichedLog net6.0 2.97μs 1.19ns 4.59ns 0.0298 0 0 2.2 KB
master EnrichedLog netcoreapp3.1 4.17μs 1.12ns 4.34ns 0.0293 0 0 2.2 KB
master EnrichedLog net472 4.8μs 1.17ns 4.53ns 0.32 0 0 2.02 KB
#6752 EnrichedLog net6.0 3.13μs 0.862ns 3.34ns 0.0314 0 0 2.2 KB
#6752 EnrichedLog netcoreapp3.1 4.29μs 1.05ns 3.92ns 0.0298 0 0 2.2 KB
#6752 EnrichedLog net472 4.97μs 1.56ns 6.03ns 0.319 0 0 2.02 KB
Benchmarks.Trace.RedisBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master SendReceive net6.0 1.35μs 0.595ns 2.31ns 0.0162 0 0 1.14 KB
master SendReceive netcoreapp3.1 1.81μs 0.881ns 3.3ns 0.0154 0 0 1.14 KB
master SendReceive net472 2.14μs 1.16ns 4.48ns 0.183 0 0 1.16 KB
#6752 SendReceive net6.0 1.36μs 0.726ns 2.81ns 0.0157 0 0 1.14 KB
#6752 SendReceive netcoreapp3.1 1.8μs 1.13ns 4.39ns 0.0146 0 0 1.14 KB
#6752 SendReceive net472 2.06μs 0.773ns 2.99ns 0.184 0 0 1.16 KB
Benchmarks.Trace.SerilogBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master EnrichedLog net6.0 2.72μs 1.73ns 6.49ns 0.0218 0 0 1.6 KB
master EnrichedLog netcoreapp3.1 3.95μs 2.2ns 8.52ns 0.0217 0 0 1.65 KB
master EnrichedLog net472 4.29μs 1.29ns 4.66ns 0.324 0 0 2.04 KB
#6752 EnrichedLog net6.0 2.74μs 0.921ns 3.57ns 0.0219 0 0 1.6 KB
#6752 EnrichedLog netcoreapp3.1 3.91μs 2.98ns 11.2ns 0.0212 0 0 1.65 KB
#6752 EnrichedLog net472 4.14μs 2.37ns 8.85ns 0.323 0 0 2.04 KB
Benchmarks.Trace.SpanBenchmark - Slower ⚠️ Same allocations ✔️

Slower ⚠️ in #6752

Benchmark diff/base Base Median (ns) Diff Median (ns) Modality
Benchmarks.Trace.SpanBenchmark.StartFinishScope‑net6.0 1.157 478.81 554.04

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master StartFinishSpan net6.0 417ns 0.11ns 0.425ns 0.00815 0 0 576 B
master StartFinishSpan netcoreapp3.1 628ns 0.892ns 3.46ns 0.00796 0 0 576 B
master StartFinishSpan net472 603ns 0.204ns 0.791ns 0.0917 0 0 578 B
master StartFinishScope net6.0 479ns 0.113ns 0.422ns 0.00981 0 0 696 B
master StartFinishScope netcoreapp3.1 724ns 0.469ns 1.75ns 0.00937 0 0 696 B
master StartFinishScope net472 865ns 0.862ns 3.23ns 0.104 0 0 658 B
#6752 StartFinishSpan net6.0 408ns 0.171ns 0.663ns 0.00803 0 0 576 B
#6752 StartFinishSpan netcoreapp3.1 569ns 2.71ns 10.5ns 0.0078 0 0 576 B
#6752 StartFinishSpan net472 583ns 0.21ns 0.812ns 0.0915 0 0 578 B
#6752 StartFinishScope net6.0 554ns 0.456ns 1.76ns 0.00983 0 0 696 B
#6752 StartFinishScope netcoreapp3.1 772ns 0.224ns 0.777ns 0.00936 0 0 696 B
#6752 StartFinishScope net472 879ns 0.364ns 1.41ns 0.104 0 0 658 B
Benchmarks.Trace.TraceAnnotationsBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch Method Toolchain Mean StdError StdDev Gen 0 Gen 1 Gen 2 Allocated
master RunOnMethodBegin net6.0 685ns 0.291ns 1.13ns 0.00995 0 0 696 B
master RunOnMethodBegin netcoreapp3.1 999ns 1.37ns 5.3ns 0.00905 0 0 696 B
master RunOnMethodBegin net472 1.01μs 0.135ns 0.506ns 0.104 0 0 658 B
#6752 RunOnMethodBegin net6.0 671ns 0.21ns 0.815ns 0.00985 0 0 696 B
#6752 RunOnMethodBegin netcoreapp3.1 989ns 0.756ns 2.93ns 0.0098 0 0 696 B
#6752 RunOnMethodBegin net472 1.07μs 0.501ns 1.94ns 0.104 0 0 658 B

@andrewlock andrewlock merged commit c7d03b1 into master Mar 11, 2025
63 of 69 checks passed
@andrewlock andrewlock deleted the andrew/ci/drop-windows-throughput-tests branch March 11, 2025 15:01
@github-actions github-actions bot added this to the vNext-v3 milestone Mar 11, 2025
andrewlock added a commit that referenced this pull request Mar 11, 2025
## Summary of changes

Remove the Profiler throughput tests on Windows (like #6752)

## Reason for change

These tests fail a _lot_ and we're hoping to migrate away soon anyway.
And @gleocadie asked me to remove them 😄

Also saw occasional timeouts in throughput tests, so bump the timeout

## Implementation details

Remove the windows job and bump timeouts throughout

## Test coverage

This is the test really
ddyurchenko pushed a commit that referenced this pull request Mar 31, 2025
## Summary of changes

Removes the Windows throughput tests

## Reason for change

They've been broken for ~3 weeks, nobody's looking at them, they slow
down the pipelines, and we're migrating to the benchmarking platform
anyway

## Implementation details

Delete the Windows job from the throughput stage - I've left all the
other infra for it, as we will likely remove _all_ of the crank stuff
post- benchmarking platform transition

## Test coverage

This is the test

## Other details

Hopefully the compare stage works, this PR will show if it doesn't
though
ddyurchenko pushed a commit that referenced this pull request Mar 31, 2025
## Summary of changes

Remove the Profiler throughput tests on Windows (like #6752)

## Reason for change

These tests fail a _lot_ and we're hoping to migrate away soon anyway.
And @gleocadie asked me to remove them 😄

Also saw occasional timeouts in throughput tests, so bump the timeout

## Implementation details

Remove the windows job and bump timeouts throughout

## Test coverage

This is the test really
chojomok pushed a commit that referenced this pull request Jul 15, 2025
## Summary of changes

Removes the Windows throughput tests

## Reason for change

They've been broken for ~3 weeks, nobody's looking at them, they slow
down the pipelines, and we're migrating to the benchmarking platform
anyway

## Implementation details

Delete the Windows job from the throughput stage - I've left all the
other infra for it, as we will likely remove _all_ of the crank stuff
post- benchmarking platform transition

## Test coverage

This is the test

## Other details

Hopefully the compare stage works, this PR will show if it doesn't
though
chojomok pushed a commit that referenced this pull request Jul 15, 2025
## Summary of changes

Remove the Profiler throughput tests on Windows (like #6752)

## Reason for change

These tests fail a _lot_ and we're hoping to migrate away soon anyway.
And @gleocadie asked me to remove them 😄

Also saw occasional timeouts in throughput tests, so bump the timeout

## Implementation details

Remove the windows job and bump timeouts throughout

## Test coverage

This is the test really
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
area:benchmarks Benchmarks, throughput tests, Crank, Bombardier, etc area:builds project files, build scripts, pipelines, versioning, releases, packages
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants