out_prometheus_exporter: apply HTTP server options

[stondo]

Summary

Propagate buffer_chunk_size and workers from the parsed HTTP server configuration
when out_prometheus_exporter creates its embedded HTTP server.

The plugin already applies idle_timeout, buffer_max_size, and max_connections.
Omitting the other two parsed options makes valid configuration ineffective,
particularly http_server.workers.

Fixes #12110

Changes

• apply http_server.buffer_chunk_size
• apply http_server.workers
• add a runtime regression test for both options

The test uses the actual Fluent Bit engine and verifies that:

configured chunk size == embedded server chunk size
embedded server workers == 2
embedded server runtime != NULL

It fails before the fix and passes afterward.

Compatibility

The default behavior is unchanged. The patch only makes explicitly configured HTTP
server options effective.

The same partial option propagation appears in out_vivo_exporter; this PR remains
scoped to the reproduced Prometheus exporter defect.

Testing

Example configuration:

pipeline:
  inputs:
    - name: fluentbit_metrics
      scrape_interval: 1

  outputs:
    - name: prometheus_exporter
      match: "*"
      host: 127.0.0.1
      port: 2021
      http_server.buffer_chunk_size: 64K
      http_server.workers: 2

Commands:

cmake -S . -B build-container-bookworm \
  -DFLB_TESTS_RUNTIME=On \
  -DFLB_TESTS_INTERNAL=On \
  -DFLB_EXAMPLES=Off \
  -DCMAKE_CXX_FLAGS=-mavx2
cmake --build build-container-bookworm -j8
ctest --test-dir build-container-bookworm \
  -R '^flb-rt-out_prometheus_exporter$' \
  --output-on-failure
ctest --test-dir build-container-bookworm \
  -R '^flb-it-http_server$' \
  --output-on-failure
FLUENT_BIT_BINARY="$PWD/build-container-bookworm/bin/fluent-bit" \
tests/integration/.venv/bin/python -m pytest \
  tests/integration/scenarios/out_prometheus_exporter/tests/test_out_prometheus_exporter_001.py \
  -q
FLUENT_BIT_BINARY="$PWD/build-container-bookworm/bin/fluent-bit" \
VALGRIND=1 VALGRIND_STRICT=1 \
  tests/integration/.venv/bin/python -m pytest \
  tests/integration/scenarios/out_prometheus_exporter/tests/test_out_prometheus_exporter_001.py \
  -q

These ran in the documented Debian 12 rootless Podman environment. The explicit AVX2
flag is a GCC 12 compatibility workaround for bundled simdutf, not part of this
change.

Results:

• full YAML-enabled build: passed
• runtime regression test: passed 10 consecutive iterations
• HTTP server internal test: passed
• normal integration: 2 passed
• strict Valgrind integration: 2 passed
• both Valgrind runs: 0 bytes live at exit, 0 errors
• CI commit-prefix validation: passed
• DCO: signed

---

Enter [N/A] in the box, if an item is not applicable to your change.

Testing

• Example configuration file for the change
• Debug/runtime output from testing the change
• Valgrind output shows no leaks or memory corruption

Packaging

• [N/A] No packaging or container-output change

Documentation

• [N/A] Existing documented options are made effective; no new option is added

Backporting

• Candidate for a separate 5.0 backport after the master change is accepted

Fluent Bit is licensed under Apache 2.0. By submitting this pull request I understand
that this code will be released under the terms of that license.

Summary by CodeRabbit

• New Features

  • Prometheus exporter HTTP servers now honor configured buffer chunk sizes and worker counts.
  • Added support for validating multi-worker exporter operation on supported platforms.

• Tests

  • Added runtime coverage for exporter HTTP server options, including buffer sizing, worker configuration, and dynamic port allocation.

[coderabbitai]

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

• @coderabbitai resume to resume automatic reviews.
• @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

• ▶️ Resume reviews
• 🔍 Trigger review

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 50e058dd-c88a-4bfb-bb10-b1a7f8bf9a32

📥 Commits

Reviewing files that changed from the base of the PR and between d55459d and 15d84f8.

📒 Files selected for processing (3)

• plugins/out_prometheus_exporter/prom_http.c
• tests/runtime/CMakeLists.txt
• tests/runtime/out_prometheus_exporter.c

🚧 Files skipped from review as they are similar to previous changes (3)

• plugins/out_prometheus_exporter/prom_http.c
• tests/runtime/CMakeLists.txt
• tests/runtime/out_prometheus_exporter.c

Included review availability: Your plan provides up to 8 included reviews per hour; 6 remain after this review.

---

📝 Walkthrough

Walkthrough

The Prometheus exporter now forwards configured HTTP buffer and worker options to its HTTP server. A runtime test selects a free port, starts the exporter with platform-specific settings, validates the server configuration, and registers the test with CMake.

Changes

Prometheus HTTP options

Layer / File(s)	Summary
Propagate HTTP server options plugins/out_prometheus_exporter/prom_http.c	prom_http_server_create() now assigns buffer_chunk_size and workers from the HTTP server configuration.
Validate runtime configuration tests/runtime/out_prometheus_exporter.c, tests/runtime/CMakeLists.txt	The runtime test uses an ephemeral loopback port, validates buffer size and worker count, checks the multi-worker runtime on non-Windows systems, and registers the test with CMake.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Merge Risk: ⚪ Minimal · up to 15d84

The change makes two explicitly configured HTTP server options effective without altering defaults, and no actionable merge-blocking risk remains after normal checks and review.

Suggested reviewers: edsiper

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 3 functions across 2 files. (1 skipped: 1 unsupported.)	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly identifies the Prometheus exporter change to apply HTTP server options.
Linked Issues check	✅ Passed	The code propagates buffer_chunk_size and workers and adds runtime coverage for the requirements in [#12110].
Out of Scope Changes check	✅ Passed	The changes are limited to option propagation and its dedicated Prometheus exporter runtime test.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4c43d302a7

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[cosmo0920]

Looks reasonable for me but I found a nitpick on this PR.

[cosmo0920]

Is this meaningful for the current Fluent Bit testing mechanism?
This is because we already enabled to run runtime test on Windows platforms here:
#12092
#12098

So, we need to run runtime tests on Windows as much as possible.

[stondo]

You are right, thanks for the pointers to #12092 and #12098. I have dropped the Windows skip and the runtime test now builds and runs on Windows too (commit a2506bb).

Changes:

• tests/runtime/CMakeLists.txt: registers flb-rt-out_prometheus_exporter unconditionally again.
• get_free_port(): calls WSAStartup() under #ifdef _WIN32 before socket(), matching the idiom in in_http.c and core-timeout.c.
• The multi-worker assertion is kept portable: shared listener ports (SO_REUSEPORT) are rejected by flb_net_socket_share_port() on Windows, so the test uses http_server.workers=2 on POSIX and 1 on Windows. The buffer_chunk_size pass-through (the actual regression from out_prometheus_exporter ignores HTTP server worker and buffer options #12110) is verified on both platforms; the runtime != NULL check is guarded to the multi-worker path only, since a single worker runs on the caller event loop.

Let me know if you would prefer a different split.

[stondo]

Follow-up after the latest push (53a16a1dc):

The original Linux/UBSAN failure this branch was addressing is now fixed. In particular, the previously relevant run-ubuntu-unit-tests (-DSANITIZE_UNDEFINED=On, 3.31.6, clang, clang++) job is green on the current head, and the other Ubuntu unit-test variants are also passing.

The remaining red jobs do not appear to be caused by this change:

• macOS TSan is failing in in_simple_systems.c:565 with Check records->num_records >= 20... failed
• Windows x64 unit tests are failing in in_syslog_udp_plaintext_expect.ps1 because the syslog input cannot bind 127.0.0.1:50003
• Windows 32-bit packaging is failing with CPack Error: Cannot find NSIS compiler makensis

Those failures are outside the out_prometheus_exporter / alignment-read path touched by this PR.

[stondo]

Rebased onto current master (d55459d49) and re-triggered CI. Head is now 15d84f880.

Dropped one commit as redundant. compat: avoid misaligned dword reads in default helper is already on master in identical form (memcpy() into a local in the non-FLB_ENFORCE_ALIGNMENT path), so the rebase skipped it. The PR is back to its two intended commits: the prom_http.c fix and the runtime test.

On the two failing jobs from the previous run, both are unrelated to this change:

• run-macos-unit-tests (-DFLB_SANITIZE_THREAD=On, 3.31.6): the failure is flb-rt-in_dummy, at in_simple_systems.c:565: Check records->num_records >= 20. That is a rate-based assertion in the dummy input runtime test, and it is timing-sensitive under ThreadSanitizer's instrumentation overhead. This PR touches only out_prometheus_exporter.
• run-windows-unit-tests (Windows 64bit, x64, ...): 171 of 172 tests passed, the single failure being in_syslog_udp_plaintext_expect.ps1. Again unrelated to the Prometheus exporter, and flb-rt-out_prometheus_exporter itself passed in that same run.

Local verification on the rebased head: flb-rt-out_prometheus_exporter passes 3/3 consecutive runs.

@cosmo0920 the Windows runtime-test point you raised is addressed as of a2506bb (carried through the rebase): the target is registered unconditionally, get_free_port() calls WSAStartup() under #ifdef _WIN32, and the worker count is 2 on POSIX / 1 on Windows since flb_net_socket_share_port() rejects SO_REUSEPORT there. The buffer_chunk_size pass-through that is the actual regression under test is exercised on both platforms.

[coderabbitai]

Note

GitHub couldn't provide a complete incremental comparison for this pull request, so CodeRabbit is performing a full review instead. This review may take a little longer.