fix--rate-limit-upsert-deadlock

[grutt]

Description

The queue loop re-upserted dynamic rate limit definitions on every tick without the per-tenant advisory lock that UpdateRateLimits uses, causing frequent deadlock detected (40P01) errors against the RateLimit table. Upserts now follow the same lock discipline and skip no-op writes entirely.

Type of change

• Bug fix (non-breaking change which fixes an issue)

What's Changed

• Acquire the per-tenant advisory lock before UpsertRateLimitsBulk, matching the lock ordering in UpdateRateLimits
• Run the upsert on the caller's optimistic transaction when present; otherwise open a short dedicated transaction
• Cache the last-written (limitValue, window) per key (5 min TTL) and skip upserts for unchanged definitions, removing per-tick write churn

Checklist

Changes have been:

• Tested (unit, integration, or manually with steps specified)
• Linted and formatted
• Documented (where applicable)
• Added to CHANGELOG (where applicable) -- see Keep a Changelog

---

🤖 AI Disclosure

• I acknowledge that an LLM was used in the creation of this Pull Request, in accordance with Hatchet's AI_POLICY.md.

• Details: Cursor (Claude) used to diagnose the deadlock.

[vercel]

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

Project	Deployment	Actions	Updated (UTC)
hatchet-docs	Ready	Preview, Comment	Jul 9, 2026 7:23pm