fix(metering): alarm when a billable meter emit is swallowed

[guangyu-reflexio]

Problem

Both metering helpers in routes/_metering.py caught every exception and logged a
warning. That is lost billable usage, and the loss is undetectable:

• the caller ignores the helper's return value, so a False goes nowhere;
• _worker_loop's own search.metering.job_failed capture can never fire, because
  these handlers swallow the exception first;
• downstream, the enterprise Sentry integration is wired at
  event_level=logging.ERROR with Sentry Logs off, so a logger.warning produces
  no event at all. Searching Sentry for one returns a false clean whether or not
  the drop is happening.

The asymmetry between the two helpers is what makes it dangerous.
_meter_applied_learnings performs a config lookup before emitting;
_meter_search_request does not. A config-store blip therefore drops
learning_applied while search_request keeps flowing, and the two meters diverge
with no signal anywhere — a shape that is easily misread as a retrieval bug rather
than a metering one.

Change

capture_anomaly("search.metering.emit_failed", level="error", ...) in both
handlers, tagged with the meter so the two are distinguishable. The logger.warning
stays; this adds the alarm rather than replacing the log.

Tests

test_metering_failure_does_not_break_search_response already drove this exact
failure path and asserted the drop happened, with a comment that it was silent.
That comment is now false, so the test asserts the alarm instead of the silence.

Mutation-verified rather than assumed: removing the applied-learnings capture fails
with assert [] == ['search.metering.emit_failed']. File restored afterwards and
confirmed by SHA-256.

• tests/server/api_endpoints/test_applied_learnings_metering.py — 24 passed
• ruff check / ruff format --check clean on both changed files

Scope

Observability only — no metering behaviour changes, and a failed emit still returns
False and still leaves the search response a 200.

Summary by CodeRabbit

• Bug Fixes

  • Improved handling of metering emission failures so requests continue successfully while failures are logged and reported for monitoring.
  • Added diagnostic reporting for failures involving applied-learning and search-request usage events.

• Tests

  • Expanded coverage to verify failure diagnostics, error severity, affected meter, and continued request success.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Essentials

Run ID: e112b5c9-5d0d-42d3-9e1f-bcf36733b924

📥 Commits

Reviewing files that changed from the base of the PR and between 3ff647c and 3c0ed74.

📒 Files selected for processing (2)

• reflexio/server/routes/_metering.py
• tests/server/api_endpoints/test_applied_learnings_metering.py

Included review availability: 1 review is currently available. Your included PR review attempts over the past 7 days set your current allowance at 4 reviews per hour.

---

📝 Walkthrough

Walkthrough

Metering helpers now report emission failures as search.metering.emit_failed anomalies with error severity, meter names, and organization IDs. Tests verify applied-learnings failure diagnostics and preserved request behavior.

Changes

Metering anomaly reporting

Layer / File(s)	Summary
Report metering emission failures reflexio/server/routes/_metering.py	Applied-learnings and search-request emission failures now call capture_anomaly with the relevant meter and organization ID while retaining error logging.
Validate applied-learnings diagnostics tests/server/api_endpoints/test_applied_learnings_metering.py	The failure test captures anomaly calls, verifies error-level search.metering.emit_failed reporting for applied_learnings, and confirms that no usage event is emitted.

Priority: ⬇️ Low

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

Merge Risk: ⚪ Minimal · up to 3c0ed

Metering emission failures now produce error-level anomaly events while preserving successful search responses and the existing best-effort metering behavior. No merge-blocking risk remains.

Suggested reviewers: yyiilluu

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 75.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 4 functions across 2 files.	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 and concisely describes the main change: reporting failed billable metering emissions instead of silently swallowing them.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

• Fix all pre-merge checks with AI

✨ Finishing Touches 💡 1

📝 Generate docstrings 💡

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/meter-emit-failure-alarm

---

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