Set .Net-specific injection package size ratchet#8392
Conversation
|
Thanks, but I set it here already 😄 #8382 |
|
Oh lovely! I missed the ping 🤦 AND the code in my worktree 🤦 🤦 Sorry for the noise, closing right now. |
Execution-Time Benchmarks Report ⏱️Execution-time results for samples comparing This PR (8392) and master. ✅ No regressions detected - check the details below Full Metrics ComparisonFakeDbCommand
HttpMessageHandler
Comparison explanationExecution-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 highlighted in **red**. The following thresholds were used for comparing the execution times:
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). Duration chartsFakeDbCommand (.NET Framework 4.8)gantt
title Execution time (ms) FakeDbCommand (.NET Framework 4.8)
dateFormat x
axisFormat %Q
todayMarker off
section Baseline
This PR (8392) - mean (76ms) : 74, 78
master - mean (72ms) : 70, 74
section Bailout
This PR (8392) - mean (80ms) : 78, 82
master - mean (76ms) : 75, 77
section CallTarget+Inlining+NGEN
This PR (8392) - mean (1,098ms) : 1058, 1139
master - mean (1,057ms) : 1027, 1087
FakeDbCommand (.NET Core 3.1)gantt
title Execution time (ms) FakeDbCommand (.NET Core 3.1)
dateFormat x
axisFormat %Q
todayMarker off
section Baseline
This PR (8392) - mean (119ms) : 115, 123
master - mean (114ms) : 111, 116
section Bailout
This PR (8392) - mean (120ms) : crit, 117, 124
master - mean (113ms) : 111, 116
section CallTarget+Inlining+NGEN
This PR (8392) - mean (810ms) : 782, 839
master - mean (776ms) : 754, 797
FakeDbCommand (.NET 6)gantt
title Execution time (ms) FakeDbCommand (.NET 6)
dateFormat x
axisFormat %Q
todayMarker off
section Baseline
This PR (8392) - mean (105ms) : 101, 109
master - mean (99ms) : 96, 102
section Bailout
This PR (8392) - mean (106ms) : crit, 103, 109
master - mean (101ms) : 98, 103
section CallTarget+Inlining+NGEN
This PR (8392) - mean (949ms) : 917, 982
master - mean (932ms) : 895, 969
FakeDbCommand (.NET 8)gantt
title Execution time (ms) FakeDbCommand (.NET 8)
dateFormat x
axisFormat %Q
todayMarker off
section Baseline
This PR (8392) - mean (105ms) : 101, 109
master - mean (98ms) : 95, 100
section Bailout
This PR (8392) - mean (106ms) : crit, 103, 108
master - mean (100ms) : 97, 102
section CallTarget+Inlining+NGEN
This PR (8392) - mean (839ms) : 805, 873
master - mean (812ms) : 774, 849
HttpMessageHandler (.NET Framework 4.8)gantt
title Execution time (ms) HttpMessageHandler (.NET Framework 4.8)
dateFormat x
axisFormat %Q
todayMarker off
section Baseline
This PR (8392) - mean (189ms) : 186, 193
master - mean (189ms) : 186, 192
section Bailout
This PR (8392) - mean (193ms) : 191, 194
master - mean (192ms) : 190, 194
section CallTarget+Inlining+NGEN
This PR (8392) - mean (1,134ms) : 1088, 1181
master - mean (1,129ms) : 1080, 1178
HttpMessageHandler (.NET Core 3.1)gantt
title Execution time (ms) HttpMessageHandler (.NET Core 3.1)
dateFormat x
axisFormat %Q
todayMarker off
section Baseline
This PR (8392) - mean (272ms) : 269, 275
master - mean (271ms) : 268, 274
section Bailout
This PR (8392) - mean (272ms) : 268, 275
master - mean (271ms) : 269, 274
section CallTarget+Inlining+NGEN
This PR (8392) - mean (919ms) : 901, 938
master - mean (915ms) : 898, 932
HttpMessageHandler (.NET 6)gantt
title Execution time (ms) HttpMessageHandler (.NET 6)
dateFormat x
axisFormat %Q
todayMarker off
section Baseline
This PR (8392) - mean (266ms) : 262, 270
master - mean (265ms) : 261, 268
section Bailout
This PR (8392) - mean (265ms) : 263, 268
master - mean (265ms) : 262, 267
section CallTarget+Inlining+NGEN
This PR (8392) - mean (1,125ms) : 1071, 1179
master - mean (1,126ms) : 1080, 1172
HttpMessageHandler (.NET 8)gantt
title Execution time (ms) HttpMessageHandler (.NET 8)
dateFormat x
axisFormat %Q
todayMarker off
section Baseline
This PR (8392) - mean (264ms) : 259, 268
master - mean (263ms) : 259, 268
section Bailout
This PR (8392) - mean (264ms) : 261, 266
master - mean (263ms) : 261, 266
section CallTarget+Inlining+NGEN
This PR (8392) - mean (1,012ms) : 979, 1046
master - mean (1,010ms) : 974, 1047
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Summary of changes
Introduce language-specific thresholds for the ratchet.
Reason for change
Due to a change PHP-side libdatadog was bumped to v29. This version bump caused PHP images to blow past their previous limit.
Simultaneously, changes to the
datadog-packagebinary were picked up. Notably thegithub.com/klauspost/compress/zstddependency was bumped from 1.17.4 to 1.18.1. This bump included in 1.17.5 a default value change for the compression context size, downscaled from 32MB to 8MB (fixing compatibility with browsers like Chrome), causing drastic changes in compression results.Unbeknownst to the second change, PHP therefore updated the global limits, attributing the size change to the libdatadog bump (which was true, but not for the whole of it):
This allowed a change to slip through to other languages, where the new
datadog-packagebinary caused regressions in package size without change to the inputs. Due to the wildly differing sizes of packages across languages the regression is silent, rendering the size threshold ratchet moot.To combat such silently creeping regressions, this PR introduces language-specific thresholds for the ratchet.
Implementation details
Overrides variables in one pipeline
Test coverage
Other details
datadog-packagewas fixed in https://github.com/DataDog/datadog-packages/pull/64 which may take some time to trickle down.