chore(profiler): better compression testing and minor fixes by felixge · Pull Request #4697 · DataDog/dd-trace-go

felixge · 2026-04-27T07:15:00Z

What does this PR do?

See commits.

Motivation

Follow-up cleanups to #4696 suggested by my clanker. Separate PR to make the main one easier to review.

Reviewer's Checklist

Changed code has unit tests for its functionality at or near 100% coverage.
System-Tests covering this feature have been added and enabled with the va.b.c-dev version tag.
There is a benchmark for any new code, or changes to existing code.
If this interacts with the agent in a new way, a system test has been added.
New code is free of linting errors. You can check this by running make lint locally.
New code doesn't break existing tests. You can check this by running make test locally.
Add an appropriate team label so this PR gets put in the right place for the release notes.
All generated files are up to date. You can check this by running make generate locally.
Non-trivial go.mod changes, e.g. adding new modules, are reviewed by @DataDog/dd-trace-go-guild. Make sure all nested modules are up to date by running make fix-modules locally.

Unsure? Have a question? Request a review!

codecov · 2026-04-27T07:19:10Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 62.04%. Comparing base (997e8b2) to head (18fcb7c).

Additional details and impacted files

Files with missing lines	Coverage Δ
profiler/compression.go	`88.33% <100.00%> (ø)`

... and 439 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

datadog-datadog-prod-us1 · 2026-04-27T07:19:16Z

🎉 All green!

❄️ No new flaky tests detected
🧪 All tests passed

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 61.40% (+3.96%)

_{This comment will be updated automatically if new data arrives.

🔗 Commit SHA: 18fcb7c | Docs | Datadog PR Page | Give us feedback!}

pr-commenter · 2026-04-27T08:23:38Z

Benchmarks

Benchmark execution time: 2026-04-29 07:56:33

Comparing candidate commit 18fcb7c in PR branch felix.geisendoerfer/strengthen-gzip-compression-tests with baseline commit 997e8b2 in branch main.

Found 0 performance improvements and 0 performance regressions! Performance is the same for 224 metrics, 0 unstable metrics.

Explanation

This is an A/B test comparing a candidate commit's performance against that of a baseline commit. Performance changes are noted in the tables below as:

🟩 = significantly better candidate vs. baseline
🟥 = significantly worse candidate vs. baseline

We compute a confidence interval (CI) over the relative difference of means between metrics from the candidate and baseline commits, considering the baseline as the reference.

If the CI is entirely outside the configured SIGNIFICANT_IMPACT_THRESHOLD (or the deprecated UNCONFIDENCE_THRESHOLD), the change is considered significant.

Feel free to reach out to #apm-benchmarking-platform on Slack if you have any questions.

More details about the CI and significant changes

You can imagine this CI as a range of values that is likely to contain the true difference of means between the candidate and baseline commits.

CIs of the difference of means are often centered around 0%, because often changes are not that big:

---------------------------------(------|---^--------)-------------------------------->
                              -0.6%    0%  0.3%     +1.2%
                                 |          |        |
         lower bound of the CI --'          |        |
sample mean (center of the CI) -------------'        |
         upper bound of the CI ----------------------'

As described above, a change is considered significant if the CI is entirely outside the configured SIGNIFICANT_IMPACT_THRESHOLD (or the deprecated UNCONFIDENCE_THRESHOLD).

For instance, for an execution time metric, this confidence interval indicates a significantly worse performance:

----------------------------------------|---------|---(---------^---------)---------->
                                       0%        1%  1.3%      2.2%      3.1%
                                                  |   |         |         |
       significant impact threshold --------------'   |         |         |
                      lower bound of CI --------------'         |         |
       sample mean (center of the CI) --------------------------'         |
                      upper bound of CI ----------------------------------'

nsrip-dd

LGTM. Confirmed the new tests fail if the fixes are reverted. I was gonna suggest b3f1228 as a followup to the parent PR so it's good to see it here.

### What does this PR do? Adds a functional gzip compression mode for the profiler via `DD_PROFILING_DEBUG_COMPRESSION_SETTINGS=gzip`. This was already kinda working, but not really, hence the "fix" title. - Defaults `gzip` to gzip-6, matching the gzip level legacy mode applies when it compresses profiler-produced profiles. - Adds gzip-to-gzip recompression so gzip mode works for runtime profiles that are already gzip-compressed at gzip-1. - Adds tests for gzip mode and gzip recompression. See #4697 for some follow-up cleanups. ### Motivation This is done for incident-53386, where zstd caused a memory usage increase. The gzip mode provides an alternative to the default zstd compression while keeping profiler uploads compressed. ### Reviewer's Checklist - [x] Changed code has unit tests for its functionality at or near 100% coverage. - [ ] [System-Tests](https://github.com/DataDog/system-tests/) covering this feature have been added and enabled with the va.b.c-dev version tag. - [ ] There is a benchmark for any new code, or changes to existing code. - [ ] If this interacts with the agent in a new way, a system test has been added. - [ ] New code is free of linting errors. You can check this by running `make lint` locally. - [x] New code doesn't break existing tests. You can check this by running `make test` locally. - [ ] Add an appropriate team label so this PR gets put in the right place for the release notes. - [ ] All generated files are up to date. You can check this by running `make generate` locally. - [ ] Non-trivial go.mod changes, e.g. adding new modules, are reviewed by @DataDog/dd-trace-go-guild. Make sure all nested modules are up to date by running `make fix-modules` locally. Tests run: - `go test ./profiler -run 'TestNewCompressionPipeline|TestDebugCompressionEnv/explicit-gzip' -count=1` Co-authored-by: felix.geisendoerfer <[email protected]>

Close gzip readers after io.Copy in the gzip recompression paths so any final decompressor error state is observed. This applies to both gzip->gzip and gzip->zstd recompression because both use a gzip reader for the input stream. This is not believed to fix any real issues, it's just done to follow the proper life cycle.

Both recompressors decompress gzip input via an io.Pipe and forward the decoded bytes to an output compressor; only the output type differed. Collapse them into a single gzipDecodingRecompressor parameterized over the compressor interface, which both *kgzip.Writer and *sharedZstdEncoder already satisfy.

If kgzip.NewReader fails (e.g. on a corrupt or non-gzip stream) the goroutine inside gzipDecodingRecompressor used to exit immediately, leaving io.Pipe with no reader. Any in-flight or subsequent pw.Write from the caller would then block forever, also preventing Close from ever draining the error channel. Close the read end of the pipe with the final error before sending it on the err channel, so pending and future writes fail fast with that error and the caller can complete normally. Includes a regression test that runs inside a testing/synctest bubble so the deadlock is detected deterministically rather than via a wall-clock timeout: synctest.Wait returns once every goroutine in the bubble is durably blocked or has exited, and the test then asserts that Write actually returned with an error.

### What does this PR do? Adds a functional gzip compression mode for the profiler via `DD_PROFILING_DEBUG_COMPRESSION_SETTINGS=gzip`. This was already kinda working, but not really, hence the "fix" title. - Defaults `gzip` to gzip-6, matching the gzip level legacy mode applies when it compresses profiler-produced profiles. - Adds gzip-to-gzip recompression so gzip mode works for runtime profiles that are already gzip-compressed at gzip-1. - Adds tests for gzip mode and gzip recompression. See #4697 for some follow-up cleanups. ### Motivation This is done for incident-53386, where zstd caused a memory usage increase. The gzip mode provides an alternative to the default zstd compression while keeping profiler uploads compressed. ### Reviewer's Checklist - [x] Changed code has unit tests for its functionality at or near 100% coverage. - [ ] [System-Tests](https://github.com/DataDog/system-tests/) covering this feature have been added and enabled with the va.b.c-dev version tag. - [ ] There is a benchmark for any new code, or changes to existing code. - [ ] If this interacts with the agent in a new way, a system test has been added. - [ ] New code is free of linting errors. You can check this by running `make lint` locally. - [x] New code doesn't break existing tests. You can check this by running `make test` locally. - [ ] Add an appropriate team label so this PR gets put in the right place for the release notes. - [ ] All generated files are up to date. You can check this by running `make generate` locally. - [ ] Non-trivial go.mod changes, e.g. adding new modules, are reviewed by @DataDog/dd-trace-go-guild. Make sure all nested modules are up to date by running `make fix-modules` locally. Tests run: - `go test ./profiler -run 'TestNewCompressionPipeline|TestDebugCompressionEnv/explicit-gzip' -count=1` Co-authored-by: felix.geisendoerfer <[email protected]>

felixge changed the title ~~chore: strengthen gzip compression tests~~ test(profiler): verify gzip compression levels Apr 27, 2026

felixge changed the title ~~test(profiler): verify gzip compression levels~~ profiler: better compression testing and minor fixes Apr 27, 2026

felixge marked this pull request as ready for review April 27, 2026 07:19

felixge requested a review from a team as a code owner April 27, 2026 07:19

felixge mentioned this pull request Apr 27, 2026

fix(profiler): support gzip compression mode #4696

Merged

9 tasks

felixge changed the title ~~profiler: better compression testing and minor fixes~~ chore(profiler): better compression testing and minor fixes Apr 27, 2026

nsrip-dd approved these changes Apr 27, 2026

View reviewed changes

Base automatically changed from felix.geisendoerfer/fix-profiler-support-gzip-compression-mode to main April 29, 2026 07:06

gh-worker-dd-devflow-36fce6 Bot added the mergequeue-status: waiting label Apr 29, 2026

felixge added 4 commits April 29, 2026 09:17

chore: strengthen gzip compression tests

18fcb7c

felixge force-pushed the felix.geisendoerfer/strengthen-gzip-compression-tests branch from 6775043 to 18fcb7c Compare April 29, 2026 07:17

gh-worker-dd-devflow-36fce6 Bot added mergequeue-status: removed mergequeue-status: waiting mergequeue-status: queued mergequeue-status: in_progress and removed mergequeue-status: waiting mergequeue-status: removed mergequeue-status: queued labels Apr 29, 2026

gh-worker-dd-mergequeue-cf854d Bot merged commit 2743f8e into main Apr 29, 2026
193 of 195 checks passed

gh-worker-dd-devflow-36fce6 Bot removed the mergequeue-status: in_progress label Apr 29, 2026

gh-worker-dd-mergequeue-cf854d Bot deleted the felix.geisendoerfer/strengthen-gzip-compression-tests branch April 29, 2026 16:37

gh-worker-dd-devflow-36fce6 Bot added the mergequeue-status: done label Apr 29, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

chore(profiler): better compression testing and minor fixes#4697

chore(profiler): better compression testing and minor fixes#4697
gh-worker-dd-mergequeue-cf854d[bot] merged 4 commits into
mainfrom
felix.geisendoerfer/strengthen-gzip-compression-tests

felixge commented Apr 27, 2026 •

edited

Loading

Uh oh!

codecov Bot commented Apr 27, 2026 •

edited

Loading

Uh oh!

datadog-datadog-prod-us1 Bot commented Apr 27, 2026 •

edited by datadog-prod-us1-6 Bot

Loading

Uh oh!

pr-commenter Bot commented Apr 27, 2026 •

edited

Loading

Explanation

More details about the CI and significant changes

Uh oh!

nsrip-dd left a comment •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

felixge commented Apr 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What does this PR do?

Motivation

Reviewer's Checklist

Uh oh!

codecov Bot commented Apr 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

datadog-datadog-prod-us1 Bot commented Apr 27, 2026 • edited by datadog-prod-us1-6 Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pr-commenter Bot commented Apr 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Benchmarks

Explanation

More details about the CI and significant changes

Uh oh!

nsrip-dd left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

felixge commented Apr 27, 2026 •

edited

Loading

codecov Bot commented Apr 27, 2026 •

edited

Loading

datadog-datadog-prod-us1 Bot commented Apr 27, 2026 •

edited by datadog-prod-us1-6 Bot

Loading

pr-commenter Bot commented Apr 27, 2026 •

edited

Loading

nsrip-dd left a comment •

edited

Loading