[dkijania]

Why

The Nightly E2E tests workflow has hit the 185-minute job timeout on all 8 jobs every night since 2026-05-06. Over the last 30 days that is ~37,700 runner-minutes, or ~152,000 billable-equivalent minutes once the macOS (10x) and Windows (2x) multipliers are applied. No nightly run has passed in the 412 runs GitHub still has on record.

Root cause of the hang (verified from the job logs):

1. In the should deploy zkApp of generated project test, zk deploy stops early with Failed to find the fee payer's account on chain and exits.
2. executeInteractiveCommand in tests/utils/cli-utils.ts waits for the Are you sure you want to send (yes/no)? prompt via waitForText from @shimkiv/cli-testing-library. That promise never settles when the text does not appear, so the test sits until the Playwright timeout.
3. playwright.config.ts sets the per-test timeout to 180 minutes, and the job timeout is 185 minutes, so the job is killed before Playwright reports anything.

What

• executeInteractiveCommand now waits for a prompt via waitForPrompt, which races the prompt against the CLI process exit (polled every 250 ms) and a 10 minute upper bound. On failure it throws with the captured stdout/stderr.
• Playwright per-test timeout lowered from 180 to 60 minutes so a stuck test fails and reports before the job is killed.
• zk deploy now prints the underlying transport/GraphQL error when the fee payer lookup request itself fails, so a failed request can be told apart from a missing account.
• Changelog entry.

Verification

• npx tsc --noEmit, prettier and eslint pass; src/lib/deploy.test.js passes (67 tests).
• The should deploy zkApp of generated project E2E test passes locally against the mocked endpoints with o1js 2.15.0 (the version installed by the project template today), including the interactive prompt path.

Open question

The Failed to find the fee payer's account on chain failure in CI is not reproduced locally with the same o1js version, so its cause is still unknown. It started the day after o1js 2.15.0 was published (2026-05-05), but that correlation is not confirmed. With this PR the nightly will fail in ~15 minutes per job instead of 185 and the new diagnostic line will show the actual lookup error. Until the lookup failure is fixed, consider disabling the nightly schedule or trimming the matrix (node 18 is end of life; the mocked macOS and Windows jobs are the bulk of the cost).

🤖 Generated with Claude Code

[codecov]

Codecov Report

❌ Patch coverage is 50.00000% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 99.93%. Comparing base (3306054) to head (2f32fd3).

Files with missing lines	Patch %	Lines
src/lib/deploy.js	50.00%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##              main     #783      +/-   ##
===========================================
- Coverage   100.00%   99.93%   -0.07%     
===========================================
  Files           11       11              
  Lines         1492     1494       +2     
  Branches       304      305       +1     
===========================================
+ Hits          1492     1493       +1     
- Misses           0        1       +1     

Flag	Coverage Δ
unittests	99.93% <50.00%> (-0.07%)	⬇️

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

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 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.