improvement(db): add session statement/lock timeouts; simplify KB doc tx

[TheodoreSpeaks]

Summary

• Add server-side lock_timeout=5s / statement_timeout=30s session defaults via connection: {...} startup params in packages/db/db.ts. Converts silent pool wedges into loud server-side cancellations.
• Workspace archival tx keeps its own ceiling via SET LOCAL statement_timeout='5min' and lock_timeout='30s' — rare admin op stays atomic without tripping the new global default.
• KB document creation (bulk + single) drops the db.transaction + SELECT 1 ... FOR UPDATE wrapper. Because KB delete is soft (deletedAt = now) the FK alone can't guard a concurrent delete, so the new insertDocumentsIfKbAlive helper does the existence check and the insert in a single statement via INSERT...SELECT...WHERE EXISTS over jsonb_to_recordset. Atomic at the MVCC snapshot — race-free, no transaction, no row lock. Side effect: removes a processDocumentTags-uses-db-inside-tx deadlock surface.

Type of Change

• Bug fix

Testing

Tested manually. bun run lint clean. bun run check:api-validation:strict passes. Vitest: workspace lifecycle 2/2, billing 29/29, knowledge 49/49, webhook trigger 17/17.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

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

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
docs	Skipped		May 14, 2026 4:24pm

[gitguardian]

⚠️ GitGuardian has uncovered 2 secrets following the scan of your pull request.

Please consider investigating the findings and remediating the incidents. Failure to do so may lead to compromising the associated services or software components.

🔎 Detected hardcoded secrets in your pull request

GitGuardian id	GitGuardian status	Secret	Commit	Filename
29606901	Triggered	Generic High Entropy Secret	a54dcbe	apps/sim/providers/utils.test.ts	View secret
32763747	Triggered	Generic Password	3e9849b	helm/sim/tests/validators_test.yaml	View secret

🛠 Guidelines to remediate hardcoded secrets

1. Understand the implications of revoking this secret by investigating where it is used in your code.
2. Replace and store your secrets safely. Learn here the best practices.
3. Revoke and rotate these secrets.
4. If possible, rewrite git history. Rewriting git history is not a trivial act. You might completely break other contributing developers' workflow and you risk accidentally deleting legitimate data.

To avoid such incidents in the future consider

• following these best practices for managing and storing secrets including API keys and other credentials
• install secret detection on pre-commit to catch secret before it leaves your machine and ease remediation.

---

^{🦉 GitGuardian detects secrets in your source code to help developers and security teams secure the modern development process. You are seeing this because you or someone else with access to this repository has authorized GitGuardian to scan your pull request.}

[cursor]

PR Summary

Medium Risk
Introduces global Postgres lock_timeout/statement_timeout defaults, which can cause previously-long or lock-waiting queries to start failing and needs validation in production workloads. Also changes knowledge-base document creation semantics by removing a transaction/lock, so knowledge_base.updatedAt is no longer updated atomically with the insert.

Overview
Adds default Postgres session guards in packages/db/db.ts (lock_timeout=5s, statement_timeout=30s) so lock waits and long-running queries fail fast instead of wedging connections.

Updates workspace archival to SET LOCAL higher per-transaction timeouts (5min statement, 30s lock) to keep this rare, large operation reliable under the new global defaults.

Simplifies knowledge-base document creation (createDocumentRecords, createSingleDocument) by dropping the explicit transaction and SELECT ... FOR UPDATE lock, relying on FK constraints/atomic inserts and reducing deadlock/pool checkout overhead.

^{Reviewed by Cursor Bugbot for commit 0ad209f. Bugbot is set up for automated code reviews on this repo. Configure here.}

[greptile-apps]

Greptile Summary

This PR adds server-side lock_timeout=5s and statement_timeout=30s session defaults to the postgres.js connection pool, converting silent long-running query hangs into loud server-side cancellations. It also refactors KB document creation to use a single INSERT...SELECT...WHERE EXISTS statement rather than an explicit transaction with a row-level lock, and overrides the new shorter defaults in the workspace archival transaction with SET LOCAL.

• packages/db/db.ts: Adds connection: { lock_timeout: 5_000, statement_timeout: 30_000 } so every connection in the pool inherits these session-level guards at connection startup.
• apps/sim/lib/workspaces/lifecycle.ts: Prepends SET LOCAL overrides at the start of the archival transaction so the 30s default doesn't abort legitimate large-workspace cleanup operations.
• apps/sim/lib/knowledge/documents/service.ts: Replaces the explicit db.transaction + SELECT ... FOR UPDATE pattern with a new insertDocumentsIfKbAlive helper that does the existence check and insert atomically in one SQL statement via INSERT...SELECT...WHERE EXISTS.

Confidence Score: 4/5

The timeout and lock changes are safe; open questions remain around the KB document creation path that were raised in previous review threads.

The db.ts and lifecycle.ts changes are straightforward and mechanically correct — session timeouts are set at startup, and the archival transaction correctly overrides them with SET LOCAL before any DML. The insertDocumentsIfKbAlive helper is a genuine improvement for combining the soft-delete guard and insert into one SQL statement. The two open issues from prior review threads (the soft-delete race window at READ COMMITTED isolation and the non-atomic updatedAt update) are still present in this revision and would benefit from being addressed before merging to a production path.

apps/sim/lib/knowledge/documents/service.ts — the document creation path still has the two open concerns from prior review threads worth re-examining before merge.

Important Files Changed

Filename	Overview
packages/db/db.ts	Adds session-level lock_timeout (5s) and statement_timeout (30s) via postgres.js connection startup params; values are correct integer milliseconds accepted by PostgreSQL.
apps/sim/lib/workspaces/lifecycle.ts	Adds SET LOCAL statement_timeout='5min' and SET LOCAL lock_timeout='30s' at the top of the archival transaction to correctly override session defaults for this rare, long-running admin operation.
apps/sim/lib/knowledge/documents/service.ts	Introduces insertDocumentsIfKbAlive using INSERT…SELECT…WHERE EXISTS to combine the soft-delete guard and insert into one statement; the updatedAt update remains outside any transaction (as flagged in existing review threads).
apps/sim/lib/workspaces/lifecycle.test.ts	Adds execute: vi.fn().mockResolvedValue([]) to the transaction mock to handle the two new SET LOCAL calls; change is correct and minimal.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[createDocumentRecords / createSingleDocument] --> B{Upfront KB check\ndb.select WHERE deletedAt IS NULL}
    B -- KB not found --> C[throw 'Knowledge base not found']
    B -- KB alive --> D[Process tag data\nprocessDocumentTags]
    D --> E[Build NewDocumentRow array]
    E --> F[insertDocumentsIfKbAlive\nINSERT...SELECT...WHERE EXISTS\ndeleted_at IS NULL]
    F -- 0 rows returned\nKB soft-deleted in window --> G[throw 'Knowledge base not found']
    F -- N rows returned --> H[Log success]
    H --> I[db.update knowledgeBase\nset updatedAt - separate statement]
    I --> J[Return document data]

Loading

_{Reviews (2): Last reviewed commit: "fix(knowledge): close soft-delete TOCTOU..." | Re-trigger Greptile}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0ad209f. Configure here.}

[TheodoreSpeaks]

@greptile review