fix: use session exec in telemetry writer

[erichare]

Summary

• Replace telemetry writer SQLModel session.execute() calls with session.exec().
• Align telemetry writer tests with the production SQLModel async session class.
• Add coverage that turns the SQLModel session.execute deprecation warning into a test failure.

Tests

• uv run pytest src/backend/tests/unit/services/telemetry_writer/test_service.py
• uv run ruff check src/backend/base/langflow/services/telemetry_writer/service.py src/backend/tests/unit/services/telemetry_writer/test_service.py
• uv run ruff format --check src/backend/base/langflow/services/telemetry_writer/service.py src/backend/tests/unit/services/telemetry_writer/test_service.py

Summary by CodeRabbit

• Bug Fixes

  • Updated telemetry writes and retention cleanup to use the supported database session method, reducing deprecation-related issues.
  • Improved handling of telemetry buffering and cleanup so existing save and prune behavior continues to work reliably.

• Tests

  • Added coverage to ensure telemetry flushing and retention do not trigger deprecated session warnings.
  • Adjusted existing test setup to match the updated session behavior.

[coderabbitai]

Walkthrough

Telemetry writer flush and retention queries now use AsyncSession.exec(...). The unit tests were updated to use SQLModel’s async session class, check for deprecated-session warnings, and align cancellation helpers with the new call path.

Changes

Telemetry writer exec migration

Layer / File(s)	Summary
Flush uses exec src/backend/base/langflow/services/telemetry_writer/service.py	_flush now uses session.exec(...) for transaction and vertex-build inserts, with transaction rows passed via params=tx_batch and dirty flow IDs still tracked per buffered row.
Retention uses exec src/backend/base/langflow/services/telemetry_writer/service.py	_run_retention_pass now uses await session.exec(...) for transaction deletion, vertex-id selection, per-vertex deletions, and the global-cap delete, while keeping the existing dirty-flow snapshot and restore flow.
Telemetry writer tests src/backend/tests/unit/services/telemetry_writer/test_service.py	The test session maker now uses SQLModel’s async session class, a new warning-based test checks _flush() and _run_retention_pass(), and cancellation/failure helpers were updated to use exec().

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• langflow-ai/langflow#13126: Updates the same telemetry writer paths to use AsyncSession.exec(...) and adds related test coverage.

Suggested reviewers

• dkaushik94

🚥 Pre-merge checks | ✅ 8 | ❌ 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%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (8 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main change: switching the telemetry writer to session.exec().
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.
Test Coverage For New Implementations	✅ Passed	Yes: src/backend/tests/unit/services/telemetry_writer/test_service.py adds a regression test for _flush/run_retention_pass and uses the SQLModel async session class; naming matches test*.py.
Test Quality And Coverage	✅ Passed	Tests are solid: real async pytest coverage against SQLite, explicit assertions on inserts/retention/failure paths, and a warning-regression check for the exec migration.
Test File Naming And Structure	✅ Passed	PASS: backend test file is named test_service.py, uses pytest fixtures/async tests, descriptive names, and structured setup/teardown with positive/negative coverage.
Excessive Mock Usage Warning	✅ Passed	Tests use a real in-memory SQLite engine/SQLModel session; only small purpose-built fakes for settings/cancellation, with no mock framework or core-logic mocking.

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

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/release-1.10.2-telemetry-session-exec

---

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.}

[github-actions]

✅ Test Coverage Advisor

No source changes detected without accompanying tests. Thanks for keeping coverage up! 🎉

Advisory check only — never blocks merge.

[coderabbitai]

🧹 Nitpick comments (1)

src/backend/tests/unit/services/telemetry_writer/test_service.py (1)

347-361: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Good deprecation guard; consider one behavioral assertion.

This test only fails if the deprecated execute() path emits a warning — it asserts nothing about the flush/retention outcome. Since it already drives both code paths against a real in-memory SQLite, a cheap assertion (e.g. the expected post-retention row counts under max_*_to_keep = 1) would also catch a regression where exec() silently no-ops or the retention query shape breaks. Note this test also implicitly exercises the vertex-build .scalars() path flagged in service.py, so confirm it actually passes in CI.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/backend/tests/unit/services/telemetry_writer/test_service.py` around
lines 347 - 361, The deprecation guard in
test_flush_and_retention_do_not_call_deprecated_session_execute is good, but the
test should also assert the flush/retention result. After calling writer._flush
and writer._run_retention_pass, add a behavioral check on the in-memory DB
state, such as the expected row counts for transactions and vertex builds when
max_transactions_to_keep and max_vertex_builds_to_keep are both 1, so the test
catches silent regressions in writer._flush, writer._run_retention_pass, or the
vertex-build scalars path.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@src/backend/tests/unit/services/telemetry_writer/test_service.py`:
- Around line 347-361: The deprecation guard in
test_flush_and_retention_do_not_call_deprecated_session_execute is good, but the
test should also assert the flush/retention result. After calling writer._flush
and writer._run_retention_pass, add a behavioral check on the in-memory DB
state, such as the expected row counts for transactions and vertex builds when
max_transactions_to_keep and max_vertex_builds_to_keep are both 1, so the test
catches silent regressions in writer._flush, writer._run_retention_pass, or the
vertex-build scalars path.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 6ad2ddbd-51b8-4904-b52d-caecd74c6826

📥 Commits

Reviewing files that changed from the base of the PR and between a66b75a and f5dd643.

📒 Files selected for processing (2)

• src/backend/base/langflow/services/telemetry_writer/service.py
• src/backend/tests/unit/services/telemetry_writer/test_service.py

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 58.57%. Comparing base (a66b75a) to head (f5dd643).
⚠️ Report is 1 commits behind head on release-1.10.2.

❌ Your project status has failed because the head coverage (54.62%) is below the target coverage (60.00%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files

@@                Coverage Diff                 @@
##           release-1.10.2   #13821      +/-   ##
==================================================
- Coverage           58.63%   58.57%   -0.07%     
==================================================
  Files                2307     2306       -1     
  Lines              220990   220169     -821     
  Branches            34290    32960    -1330     
==================================================
- Hits               129574   128956     -618     
+ Misses              89933    89729     -204     
- Partials             1483     1484       +1     

Flag	Coverage Δ
backend	65.87% <100.00%> (-0.04%)	⬇️
frontend	57.68% <ø> (-0.09%)	⬇️
lfx	54.62% <ø> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
...base/langflow/services/telemetry_writer/service.py	80.61% <100.00%> (ø)

... and 267 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.

[github-actions]

Frontend Unit Test Coverage Report

Coverage Summary

Lines	Statements	Branches	Functions
	43.54% (58072/133362)	69.2% (7884/11392)	41.56% (1298/3123)

Unit Test Results

Tests	Skipped	Failures	Errors	Time
4978	0 💤	0 ❌	0 🔥	12m 30s ⏱️