ci: skip e2e_test on master pushes, run a daily nightly canary instead

[davidfirst]

Why

The e2e_test run on every master push gates nothing: bit_merge requires only generate_and_check_types, and the merge queue's "settled" check watches bit_merge alone (.github/scripts/merge-queue.js). Every PR already ran the full suite to get merged, so the per-merge master run was a purely informational duplicate at ~5.5k credits per merge (~440k credits/month at current merge volume).

Change

• e2e_test in build_and_test now ignores the master branch.
• A daily e2e_test canary runs in the existing nightly workflow (catches merge-order interactions that individual PR runs can't).

Extracted from #10597 (closed) — this is the single highest-value change from that PR, with none of the resource-class/parallelism changes.

[qodo-free-for-open-source-projects]

PR Summary by Qodo

Reduce master E2E runs with a daily nightly canary

⚙️ Configuration changes ✨ Enhancement 🕐 Less than 10 minutes

AI Description

• Skip redundant E2E runs on master pushes to reduce CircleCI credit consumption.
• Run E2E daily on master to detect merge-order integration regressions.
• Preserve existing PR E2E coverage and master merge gating.

Diagram

graph TD
  A["Code Push"] --> B["Build Workflow"] --> C{"Master Branch?"}
  C -- "No" --> D["E2E Test"]
  C -- "Yes" --> E["Bit Merge"]
  F["Daily Schedule"] --> G["Nightly Workflow"] --> D

Loading

High-Level Assessment

The following are alternative approaches to this PR:

1. Retain per-push master E2E

• ➕ Provides immediate feedback after every merge
• ➕ Maximizes detection frequency for merge-order regressions
• ➖ Consumes substantial CircleCI credits
• ➖ Duplicates the full suite already required before merge
• ➖ Does not participate in master merge gating

2. Use path-filtered master E2E

• ➕ Runs more frequently when likely relevant files change
• ➕ Reduces execution compared with unconditional master runs
• ➖ Adds maintenance complexity to CI configuration
• ➖ Can miss cross-cutting or merge-order interactions
• ➖ Requires reliable ownership of E2E-relevant paths

Recommendation: The proposed daily canary is the best balance: it removes an expensive, non-gating duplicate while preserving broad master integration coverage. Retaining per-push execution is unnecessarily costly, and path filtering would be more complex and less reliable for detecting cross-cutting regressions.

Files changed (1) +13 / -0

Other (1) +13 / -0

config.ymlMove master E2E coverage to the nightly workflow +13/-0

Move master E2E coverage to the nightly workflow

• Adds a master branch exclusion to the regular build_and_test E2E job while leaving non-master coverage intact. Adds the same E2E job after setup_harmony in the existing daily master-only nightly workflow, with comments documenting the gating and cost rationale.

.circleci/config.yml

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

Great, no issues found!

Qodo reviewed your code and found no material issues that require review

Tip of the day

💡 Did you know, you can show, collapse, or hide each part of a finding: code, evidence, and all

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗