Isolate Cloud AI local-file offloads from default executor

[Copilot]

Canonical issue

Linked by system.

Outcome

Prevents stalled local file reads (e.g., remote-backed mounts) in Cloud AI image paths from consuming default asyncio executor capacity shared by unrelated subsystems.

• Dedicated blocking I/O lane: added a Cloud AI–scoped thread pool + run_blocking(...) helper.
• Targeted migration (AC1): moved local _read_file_bytes offloads in AWS Rekognition, Azure Vision, and Google Cloud providers to the dedicated pool.
• Saturation observability (AC3): emits structured warning when active+queued work reaches pool capacity.
• Isolation proof (AC2): regression test saturates the Cloud AI I/O pool and verifies unrelated default asyncio.to_thread(...) work still completes.

# before
await asyncio.to_thread(_read_file_bytes, str(safe_path))

# after
await run_blocking(_read_file_bytes, str(safe_path))

Scope

• Included:
  • New module: integrations/cloud_ai/blocking_io.py (ThreadPoolExecutor + run_blocking)
  • Provider routing updates in:
    • providers/aws_rekognition.py
    • providers/azure_vision.py
    • providers/google_cloud.py
  • Regression/contract tests:
    • tests/unit/test_cloud_ai_blocking_io.py
    • provider routing assertions in each provider test file
• Explicitly excluded:
  • Repo-wide migration of all to_thread / run_in_executor(None, ...) call sites
  • Behavioral changes to non-local-file provider operations

Risk

• Risk level: medium
• Failure mode:
  • Mis-sized Cloud AI pool can increase latency for Cloud AI local-file reads under burst.
  • Invalid CLOUD_AI_IO_MAX_WORKERS value falls back to default with warning.
• Rollback:
  • Revert provider call sites to asyncio.to_thread(...) and remove blocking_io.py.

Verification

List exact automated and manual checks, tied to the current head SHA.

Head SHA: e551419

• Focused tests
• Required CI
• Review threads resolved

Production evidence

Not applicable: backend concurrency/isolation change with unit-level regression coverage; no UI or deploy-surface delta.

Agent handoff

• One canonical issue is linked
• No competing PR implements the same issue
• Acceptance criteria are satisfied
• Required checks pass on the current head
• Human decision is requested only for product, security, irreversible infrastructure, or production approval

• Fixes Blocking I/O offloads share the default executor, so one stalled read can starve the process #1234

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
v0-uvai	Ready	Preview, v0	Sep 12, 2026 9:35am UTC

[vercel]

_IO_QUEUED counter permanently leaks by 1 whenever a queued blocking-I/O work item is cancelled (via timeout or task cancellation) before it starts running.

[coderabbitai]

Important

Review skipped

Bot user detected.

To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Advanced

Run ID: bf04d445-d71d-4732-ac4e-1e1436c90c5d

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

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]

Dependency Review

✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.

Snapshot Warnings

⚠️: No snapshots were found for the head SHA a8be54f.

Ensure that dependencies are being submitted on PR branches and consider enabling retry-on-snapshot-warnings. See the documentation for more information and troubleshooting advice.

Scanned Files

None

[github-actions]

🔍 PR Validation

⚠️ PR title should follow conventional commits format