[Metricbeat] Bump elastic-agent-system-metrics for zswap and cgroups v2 CPU CFS data

[orestisfl]

Proposed commit message

Bumps the elastic-agent-system-metrics dependency to v0.14.3, v0.14.2 also included.

What's Changed

• Add Zswap Metrics by @orestisfl in Add Zswap Metrics elastic-agent-system-metrics#279
• PULL_REQUEST_TEMPLATE: Remove reference to non-existant CHANGELOG.md by @orestisfl in PULL_REQUEST_TEMPLATE: Remove reference to non-existant CHANGELOG.md elastic-agent-system-metrics#286
• fix: remove shared context timeout from matrix tests by @orestisfl in fix: remove shared context timeout from matrix tests elastic-agent-system-metrics#285
• Enable modernize linter and apply suggestions by @orestisfl in Enable modernize linter and apply suggestions elastic-agent-system-metrics#287
• Use json omitzero for struct-typed fields in cgroup structs by @orestisfl in Use json omitzero for struct-typed fields in cgroup structs elastic-agent-system-metrics#288
• [cgv2] Add CPU CFS quota, period, and weight metrics by @orestisfl in [cgv2] Add CPU CFS quota, period, and weight metrics elastic-agent-system-metrics#281
• Add missing zswap debug metrics by @orestisfl in Add missing zswap debug metrics elastic-agent-system-metrics#289

Checklist

• My code follows the style guidelines of this project
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have made corresponding change to the default configuration files
• I have added tests that prove my fix is effective or that my feature works. Where relevant, I have used the stresstest.sh script to run them under stress conditions and race detector to verify their stability.
• I have added an entry in ./changelog/fragments using the changelog tool.

Disruptive User Impact

None. New features.

1. Build metricbeat from the branch:

cd metricbeat
mage build

2. Create a test config file metricbeat.test.yml:

metricbeat.modules:
  - module: system
    period: 5s
    metricsets:
      - memory
      - process
    process.include_top_n:
      by_cpu: 5
      by_memory: 5

output.console:
  pretty: true

logging.level: debug

3. Run metricbeat:

./metricbeat -e -c metricbeat.test.yml

4. Verify zswap memory metrics appear in the output (requires Linux with zswap enabled -- check with cat /sys/module/zswap/parameters/enabled):

./metricbeat -e -c metricbeat.test.yml 2>/dev/null | jq 'select(.metricset.name == "memory") | .system.memory.zswap'

Expected output (values will vary):

{
  "compressed": 950353920,
  "uncompressed": 3813502976
}

5. To also verify the zswap.debug.* fields (requires debugfs access), grant the binary the cap_dac_read_search capability:

sudo setcap cap_dac_read_search=+ep metricbeat

Then re-run. Expected output with debug fields:

{
  "compressed": 950353920,
  "uncompressed": 3813502976,
  "debug": {
    "decompress_fail": 0,
    "pool_limit_hit": 0,
    "pool_total_size": 950353920,
    "reject_alloc_fail": 0,
    "reject_compress_fail": 0,
    "reject_compress_poor": 0,
    "reject_kmemcache_fail": 0,
    "reject_reclaim_fail": 0,
    "stored_incompressible_pages": 16779,
    "stored_pages": 931031,
    "written_back_pages": 29111
  }
}

6. Verify cfs.weight appears in process cgroup metrics (requires cgroupsv2):

./metricbeat -e -c metricbeat.test.yml 2>/dev/null | jq 'select(.metricset.name == "process") | .system.process.cgroup.cpu.cfs.weight // empty'

7. Run unit tests:

cd metricbeat
go test ./module/system/...

Related issues

• Closes [Metricbeat] Zswap Metrics #47605
• Closes system.process dataset should support cgroupv2 CPU metrics #47708

[github-actions]

🤖 GitHub comments

Just comment with:

• run docs-build : Re-trigger the docs validation. (use unformatted text in the comment!)

[github-actions]

Vale Linting Results

Summary: 1 suggestion found

💡 Suggestions (1)

File	Line	Rule	Message
docs/reference/metricbeat/exported-fields-system.md	810	Elastic.WordChoice	Consider using 'select, press, visits' instead of 'hit', unless the term is in the UI.

---

The Vale linter checks documentation changes against the Elastic Docs style guide.

To use Vale locally or report issues, refer to Elastic style guide for Vale.

[github-actions]

🔍 Preview links for changed docs

• docs/reference/metricbeat/exported-fields-system.md

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 9e3383ff-676c-4630-8fe5-dcbbbc290d33

📥 Commits

Reviewing files that changed from the base of the PR and between ff149f5 and 6eed8f8.

⛔ Files ignored due to path filters (1)

• go.sum is excluded by !**/*.sum

📒 Files selected for processing (9)

• NOTICE.txt
• changelog/fragments/1772011445-zswap-metrics.yaml
• changelog/fragments/1772636540-zswap-metrics.yaml
• docs/reference/metricbeat/exported-fields-system.md
• go.mod
• metricbeat/module/system/fields.go
• metricbeat/module/system/memory/_meta/fields.yml
• metricbeat/module/system/process/_meta/fields.yml
• metricbeat/module/system/test_system.py

---

📝 Walkthrough

Walkthrough

This pull request adds zswap memory metrics collection to the system memory metricset and introduces cgroupv2 CPU weight metric support to the system.process dataset. Changes include updating the elastic-agent-system-metrics dependency from v0.14.1 to v0.14.3 in go.mod and NOTICE.txt, adding field definitions for zswap compression statistics and debug counters under memory.swap, adding the CPU weight field for cgroup v2 CPU scheduling, updating documentation with the new metric field definitions, and modifying test fixtures to account for the new Linux-specific zswap fields.

🚥 Pre-merge checks | ✅ 2

✅ Passed checks (2 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	The PR successfully addresses both linked issues: #47605 (zswap metrics) and #47708 (cgroupv2 CPU metrics). All required metrics from /proc/meminfo, /proc/vmstat, and /sys/kernel/debug/zswap are documented; cgroup v2 CPU weight field is added.
Out of Scope Changes check	✅ Passed	All changes are directly scoped to the two linked issues: dependency bump to elastic-agent-system-metrics v0.14.3, documentation updates for zswap and CPU metrics, field definitions, and test updates. No extraneous changes detected.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

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

[pierrehilbert]

/test

[mauri870]

Just a curiosity: I assume this metrics apply to EEVDF as well, now that it has been the default Linux scheduler since v6.6. It just feels a bit weird that we still have cfs in the metric name.

[vishaangelova]

LGTM, left some small suggestions for your consideration.

[vishaangelova]

Suggested change

	CPU weight for the cgroup (cgroupv2). Used by the CFS
	scheduler to determine the share of CPU time available
	to the cgroup. Range is 1 to 10000, default 100.
	CPU weight for the cgroup (cgroupv2). Used by the CFS
	scheduler to determine the share of CPU time available
	to the cgroup. Valid values range from 1 to 10000. The
	default value is 100.

[vishaangelova]

Suggested change

	zswap compressed swap cache metrics. Available on Linux when zswap is enabled.
	Metrics for the zswap compressed swap cache. Available on Linux when zswap is enabled.

[rdner]

@orestisfl we should be able to bump github.com/elastic/elastic-agent-system-metrics to v0.14.3 on all active version branches.

Should we backport this PR? Or you planned to address it differently?