Add benchmarks for DABs #4194
Merged
Add benchmarks for DABs #4194
Conversation
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
denik
requested review from
andrewnester,
anton-107,
pietern and
shreyas-goenka
as code owners
Contributor
Author
|
Current output: Same but with GOGC=off: |
Collaborator
|
Commit: f4995ea
22 interesting tests: 21 KNOWN, 1 SKIP
Top 21 slowest tests (at least 2 minutes):
|
Contributor
shreyas-goenka
left a comment
shreyas-goenka
left a comment
There was a problem hiding this comment.
Looks good, just some minor comments.
|
|
||
| Benchmarks are regular acceptance test that log measurements in certain format. The output can be fed to `tools/bench_parse.py` to print a summary table. | ||
|
|
||
| Test runner recognizes benchmark as having "benchmark" anywhere in the path. For these tests parallel execution is disabled if and only if BENCHMARK\_PARAMS variable is set. |
Contributor
There was a problem hiding this comment.
Should we have a separate top level benchmarks directory? That's better for discoverability since we can eyeball all benchmarks that exists. That also maintains a clear separation between tests and benchmarks.
Contributor
Author
There was a problem hiding this comment.
It would be harder to separate different projects inside CLI if we add another entry point. Today we have roughly this structure
- cmd/project
- acceptance/project
and we have tools that depend on it (testmask). Adding top level benchmarks directory requires making those tools aware.
cc @pietern
Contributor
Author
There was a problem hiding this comment.
note, this is something we can revisit later.
| @@ -0,0 +1,2 @@ | |||
|
|
|||
| >>> benchmark.py $CLI bundle plan > /dev/null | |||
Contributor
There was a problem hiding this comment.
Can we also log / commit the benchmarks numbers? For local vs cloud? Useful numbers to eyeball and keep track of.
Contributor
Author
There was a problem hiding this comment.
I logged them as a comment above. Absolute numbers are not useful due to different machines. I expect when someone improves the benchmarks they will post before/after numbers.
| test_match = re.match(r"=== RUN\s+(.+)", line) | ||
| if test_match: | ||
| current_test = test_match.group(1) | ||
| current_test = current_test.removeprefix("TestAccept/bundle/benchmarks/") |
Contributor
There was a problem hiding this comment.
why is this hardcoded here?
shreyas-goenka
approved these changes
Jan 6, 2026
Collaborator
|
Commit: 9747a93
26 interesting tests: 21 KNOWN, 2 flaky, 2 FAIL, 1 SKIP
Top 50 slowest tests (at least 2 minutes):
|
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.
Changes
Benchmarks are regular acceptance test that log measurements in certain format. The output can be fed to "tools/bench_parse.py" to print a summary table.
Test runner recognizes benchmark as having “benchmark” anywhere in the path. For these tests parallel execution is disabled if and only if BENCHMARK_PARAMS variable is set.
The benchmarks make use of two scripts:
gen_config.py —jobs Nto generate a config with N jobsbenchmark.py commandto run command a few times and log the time measurements.The default number of runs in benchmark.py depends on BENCHMARK_PARAMS variable. If it’s set, the default number is 5. Otherwise it is 1.
Note, bundle/benchmarks/deploy benchmarks the case when there are no changes to be deployed. If we actually measure first deployment, the time goes up from 40s to 140s for terraform but time for direct is more or less the same.
Why
Understand choke points, set the baseline for further optimizations.
Tests
Each benchmark by default is run as regular acceptance test (with a single run).