DRIVERS-3254: Add perfcomp to drivers-evergreen-tools

[zhouselena]

DRIVERS-3254

Move perfcomp from Go Driver to DET.

• perfcomp CLI, compare calculates the information, mdreport generates markdown comment
• Added project-to-perf-context map to set up compatibility with all drivers
• reference: GODRIVER-3102 where the tool was initially created

[blink1073]

So far so good, I ran it against a patch build:

./bin/perfcomp compare --project mongo-python-driver 68967120d8183c0007458ce2
2025/08/11 12:14:54 Successfully connected to MongoDB Analytics node.
2025/08/11 12:14:54 Successfully retrieved 0 docs from version 68967120d8183c0007458ce2.
2025/08/11 12:14:54 🧪 Performance Results
2025/08/11 12:14:54 There were no significant changes to the performance to report for version 68967120d8183c0007458ce2.
| Benchmark | Measurement | % Change | Patch Value | Stable Region | H-Score | Z-Score |
| --------- | ----------- | -------- | ----------- | ------------- | ------- | ------- |

*For a comprehensive view of all microbenchmark results for this PR's commit, please check out the Evergreen perf task for this patch.*

I'll see if I can find a build that does have a change point.

[blink1073]

I'd also suggest adding a run.sh since it wasn't immediately obvious how to run the command from the bin dir.

[zhouselena]

I'd also suggest adding a run.sh since it wasn't immediately obvious how to run the command from the bin dir.

@blink1073 Added .evergreen/run-perf-comp.sh which builds and runs compare, & added instructions to the readme for clarity! Also I fixed up the formatting when there are no significant tests so it shouldn't spit out an empty table. Thanks for testing it out!

[blink1073]

I purposely caused a slowdown in bson encoding/decoding, and verified that the benchmark went from bout 100 to 7 Mbps for DeepEncoding, but did not get a result.

$ ./bin/perfcomp compare --project mongo-python-driver 689df49066d4460007dc4961
2025/08/14 10:46:13 Successfully connected to MongoDB Analytics node.
2025/08/14 10:46:13 Successfully retrieved 0 docs from version 689df49066d4460007dc4961.
2025/08/14 10:46:14 🧪 Performance Results
2025/08/14 10:46:14 There were no significant changes to the performance to report for version 689df49066d4460007dc4961.
| Benchmark | Measurement | % Change | Patch Value | Stable Region | H-Score | Z-Score |
| --------- | ----------- | -------- | ----------- | ------------- | ------- | ------- |

*For a comprehensive view of all microbenchmark results for this PR's commit, please check out the Evergreen perf task for this patch.*

[zhouselena]

@blink1073 did you create a performance context for the python driver? this is needed to filter for all the python project data, it might also take a few hours for the data to be tagged. If not I can also create the context and test it out for python driver!

[blink1073]

Okay, I'll wait and try later. I targeted this context: https://performance-monitoring-and-analysis.server-tig.prod.corp.mongodb.com/context/name/PyMongo%20Perf.

[zhouselena]

Okay, I'll wait and try later. I targeted this context: https://performance-monitoring-and-analysis.server-tig.prod.corp.mongodb.com/context/name/PyMongo%20Perf.

ahh I believe thats a time series context, the performance context needs to be a triage context specifically. sorry for not specifying! if you could target a triage context that already exists it should work, if you make a new triage context that may take a few hours. I'll add that to the README for clarity

[blink1073]

Okay, I created https://performance-monitoring-and-analysis.server-tig.prod.corp.mongodb.com/context/name/PyMongo%20triage%20context. I'll try that one later. 😄

[blink1073]

Looking good for Python!

$ ./bin/perfcomp compare --perf-context "PyMongo triage context" --project mongo-python-driver --task perf-8.0-standalone --variant performance-benchmarks 689df49066d4460007dc4961
2025/08/18 10:29:41 Successfully connected to MongoDB Analytics node.
2025/08/18 10:29:41 Successfully retrieved 34 docs from version 689df49066d4460007dc4961.
2025/08/18 10:29:44 🧪 Performance Results
2025/08/18 10:29:44 The following benchmark tests for version 689df49066d4460007dc4961 had statistically significant changes (i.e., |z-score| > 1.96):

| Benchmark                      | Measurement       | % Change | Patch Value | Stable Region                                   | H-Score | Z-Score  |
| ---------                      | -----------       | -------- | ----------- | -------------                                   | ------- | -------  |
| DeepDecoding                   | megabytes_per_sec | -93.0001 | 7.1324      | Avg: 101.8930, Med: 101.6913, Stdev: 0.9885     | 0.9952  | -95.8661 |

[blink1073]

LGTM, great work!

[prestonvasquez]

Great work 👍