[waleedlatif1]

Summary

• A warm workbench run_code call has a fixed cost of about 1 s, and today the only thing logged is the total executionTime. This PR records wall time for each phase (resolve, access, lease with its lookup and cli parts, prepare, run, collect, release) and attaches it as phaseMs to the existing E2B shell/JS/Py sandbox log lines, together with sandboxSession (created/reused). The JS and Py lines also get executionTime, which the shell line already had
• No new log lines. Fields are added to existing lines only. No behavior change: the timer only reads performance.now() around awaits that were already there
• phaseMs is used only on the log line. The route destructures it and never puts it in a response
• No overhead removed in this PR. I benchmarked a warm call's steps against a real E2B sandbox (6 warm iterations, about 45 ms RTT). The floor is about 8 serial E2B round trips, and each one has a job: list and connect, the deployment CLI integrity check (launch plus stop receipt, about 250 ms, roughly a third of the floor), the output-dir sentinel, the user command (launch plus stop receipt), and the output listing. On top of those are the ownership claim/settle writes for each tracked command. None of these is redundant without a change to the durable process ownership or CLI repair design, so this PR only measures. The per-phase fields will show whether cli is the right next target in production

Type of Change

• Improvement (observability)

Testing

• New test in session-sandbox.test.ts: a warm shell call with slowed lookup, CLI check, and run reports each delay in its own phase. Fails on the pre-change source, passes after
• bun run --cwd apps/sim test on lib/execution/remote-sandbox, lib/function-execution, lib/uploads/documents, lib/workspace-files: 90 files, 1,406 tests pass
• check:comment-hygiene, check:file-names, check:explicit-any pass locally. CI runs the full gate

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing (new tests pass the test-audit authoring gate)
• 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
docs	Skipped		Nov 3, 2026 12:24am UTC

[greptile-apps]

Confidence Score: 5/5

[Low impact] The PR appears safe to merge; both previous findings are addressed and no new actionable issues were found.

Summary

Adds per-phase sandbox timings to existing execution logs without adding them to response fields.

• Sandbox logs show how long each execution phase takes.

_{Reviews (2) · Last reviewed commit: "test(sandbox): time the unchanged-CLI wa..." · Reviewed by Greptile}

[cubic-dev-ai]

No issues found across 6 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

No issues found across 6 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}