ci: reduce PR feedback loop with targeted caching#1294
Open
jeremyeder wants to merge 6 commits intoambient-code:mainfrom
Open
ci: reduce PR feedback loop with targeted caching#1294jeremyeder wants to merge 6 commits intoambient-code:mainfrom
jeremyeder wants to merge 6 commits intoambient-code:mainfrom
Conversation
Targeted caching wins across all PR workflows plus E2E image reuse via shared GHA BuildKit cache scopes. Goal: cut PR wall-clock from ~10.4m P50 to ~5-7m. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
4 tasks: E2E Docker layer caching (shared scopes from components-build), kind binary caching, golangci-lint consolidation, junit2html via pipx. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The test build tag is a superset: linting with --build-tags=test covers all production files plus test-tagged files. Reduces CI runtime by eliminating redundant pass. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace monolithic shell script with individual docker/build-push-action
steps. Each component build now reads GHA cache from components-build-deploy
workflow (scope: {component}-amd64) and writes to e2e-specific scope
(scope: e2e-{component}) for future runs.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add workflow-level KIND_VERSION env var and use actions/cache@v4 to store the kind binary in ~/k8s-tools/kind. Cache key includes OS and version for invalidation on upgrades. Avoids redundant downloads on every workflow run. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Contributor
📝 WalkthroughWalkthroughThis PR implements CI workflow optimizations to reduce PR feedback latency by refactoring E2E image builds with docker/build-push-action and caching, consolidating golangci-lint passes, caching the kind binary, and replacing pip-based junit2html installation with pipx across three core workflows. Changes
🚥 Pre-merge checks | ✅ 6✅ Passed checks (6 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
⚔️ Resolve merge conflicts
✨ Simplify code
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 2
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
.github/workflows/lint.yml (1)
144-149:⚠️ Potential issue | 🟠 MajorBackend lint skips production-only files with
--build-tags=test.
components/backend/handlers/k8s_clients_for_request_prod.gohas//go:build !testand won't be linted by the current invocation. Add a second pass without build-tag restriction to catch production-only code.Proposed fix
- name: Run golangci-lint (all build tags) uses: golangci/golangci-lint-action@v9 with: version: latest working-directory: components/backend args: --timeout=5m --build-tags=test + - name: Run golangci-lint (default build tags) + uses: golangci/golangci-lint-action@v9 + with: + version: latest + working-directory: components/backend + args: --timeout=5m🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/lint.yml around lines 144 - 149, CI currently runs golangci-lint only with --build-tags=test, which skips production-only files like components/backend/handlers/k8s_clients_for_request_prod.go (//go:build !test); add a second golangci-lint step that also runs in the same working-directory (components/backend) but without --build-tags (or with an explicit empty/omitted build-tags arg) so production-only files are linted too, keeping the existing step for test-tagged checks; target the existing action block name "Run golangci-lint (all build tags)" or add a new step named e.g. "Run golangci-lint (no build tags)" to make this explicit.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In @.github/workflows/e2e.yml:
- Around line 12-17: Add an explicit permissions block at the top-level of this
workflow (near the existing env/KIND_VERSION and concurrency keys) that scopes
GITHUB_TOKEN to only the least-privilege scopes required for the job (do NOT
rely on default write); for example set only the specific permissions you need
such as contents: read, actions: read, id-token: write, or packages: write as
appropriate for your cache/action usage, and remove any broader write defaults.
Also pin every third-party action used in the workflow to an exact commit SHA
(instead of floating tags) so the run is hermetic and auditable.
- Around line 95-96: Replace all mutable action tags with immutable commit SHAs
for each action usage (e.g., replace docker/build-push-action@v7 and the other
`@v6/`@v4/@v3 occurrences with their corresponding full commit SHAs) so every
"uses:" entry is pinned; update the 12 action instances referenced (including
docker/build-push-action and the other actions in the file) to their exact SHA
values. Also add a top-level permissions: block to the workflow that scopes the
GITHUB_TOKEN to the minimal required permissions for this workflow (declare only
the specific permission keys needed, e.g., read/write for specific resources
used), ensuring the token is not granted broad defaults. Ensure changes touch
the workflow root (top-level) and every "uses:" line that currently has a
version tag.
---
Outside diff comments:
In @.github/workflows/lint.yml:
- Around line 144-149: CI currently runs golangci-lint only with
--build-tags=test, which skips production-only files like
components/backend/handlers/k8s_clients_for_request_prod.go (//go:build !test);
add a second golangci-lint step that also runs in the same working-directory
(components/backend) but without --build-tags (or with an explicit empty/omitted
build-tags arg) so production-only files are linted too, keeping the existing
step for test-tagged checks; target the existing action block name "Run
golangci-lint (all build tags)" or add a new step named e.g. "Run golangci-lint
(no build tags)" to make this explicit.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: a3e90040-f7a9-491d-bf5b-3d9c35e7ab58
📒 Files selected for processing (5)
.github/workflows/e2e.yml.github/workflows/lint.yml.github/workflows/unit-tests.ymldocs/superpowers/plans/2026-04-11-ci-improvements.mddocs/superpowers/specs/2026-04-11-ci-improvements-design.md
Comment on lines
+12
to
17
| env: | ||
| KIND_VERSION: "v0.27.0" | ||
|
|
||
| concurrency: | ||
| group: e2e-tests-${{ github.event.pull_request.number || github.sha }} | ||
| cancel-in-progress: true |
Contributor
There was a problem hiding this comment.
Add explicit least-privilege permissions for this workflow.
The workflow introduces cache writes and multiple third-party actions but does not scope GITHUB_TOKEN permissions explicitly.
Suggested hardening baseline
env:
KIND_VERSION: "v0.27.0"
+permissions:
+ contents: read
+ actions: write
+
concurrency:
group: e2e-tests-${{ github.event.pull_request.number || github.sha }}
cancel-in-progress: trueAs per coding guidelines, .github/workflows/**/*.{yml,yaml}: - Pin action versions to SHA. Verify secrets are not exposed and permissions are scoped.
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| env: | |
| KIND_VERSION: "v0.27.0" | |
| concurrency: | |
| group: e2e-tests-${{ github.event.pull_request.number || github.sha }} | |
| cancel-in-progress: true | |
| env: | |
| KIND_VERSION: "v0.27.0" | |
| permissions: | |
| contents: read | |
| actions: write | |
| concurrency: | |
| group: e2e-tests-${{ github.event.pull_request.number || github.sha }} | |
| cancel-in-progress: true |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In @.github/workflows/e2e.yml around lines 12 - 17, Add an explicit permissions
block at the top-level of this workflow (near the existing env/KIND_VERSION and
concurrency keys) that scopes GITHUB_TOKEN to only the least-privilege scopes
required for the job (do NOT rely on default write); for example set only the
specific permissions you need such as contents: read, actions: read, id-token:
write, or packages: write as appropriate for your cache/action usage, and remove
any broader write defaults. Also pin every third-party action used in the
workflow to an exact commit SHA (instead of floating tags) so the run is
hermetic and auditable.
Comment on lines
+95
to
+96
| uses: docker/build-push-action@v7 | ||
| with: |
Contributor
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# List tag-pinned actions in the changed workflow.
rg -nP '^\s*uses:\s*[^@\s]+@v[0-9]+' .github/workflows/e2e.ymlRepository: ambient-code/platform
Length of output: 551
🏁 Script executed:
head -20 .github/workflows/e2e.ymlRepository: ambient-code/platform
Length of output: 405
🏁 Script executed:
sed -n '90,100p' .github/workflows/e2e.ymlRepository: ambient-code/platform
Length of output: 422
🏁 Script executed:
sed -n '110,120p' .github/workflows/e2e.ymlRepository: ambient-code/platform
Length of output: 500
🏁 Script executed:
sed -n '1,40p' .github/workflows/e2e.ymlRepository: ambient-code/platform
Length of output: 1194
🏁 Script executed:
sed -n '170,180p' .github/workflows/e2e.ymlRepository: ambient-code/platform
Length of output: 335
🏁 Script executed:
sed -n '230,245p' .github/workflows/e2e.ymlRepository: ambient-code/platform
Length of output: 584
🏁 Script executed:
grep -n 'permissions:' .github/workflows/e2e.ymlRepository: ambient-code/platform
Length of output: 47
Workflow actions must be pinned to commit SHAs and permissions scoped.
All action usages rely on mutable version tags (@v6, @v7, @v4, @v3) instead of immutable commit SHAs. This affects 12 action instances: lines 32, 37, 60, 78, 89, 95, 114, 133, 152, 173, 232, and 241.
Additionally, the workflow lacks an explicit top-level permissions: block to scope token access.
Replace version tags with full commit SHAs (e.g., uses: docker/build-push-action@<full-sha>) and add a top-level permissions: block specifying minimal required scopes per the guideline: .github/workflows/**/*.{yml,yaml} must pin action versions to SHA and scope permissions.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In @.github/workflows/e2e.yml around lines 95 - 96, Replace all mutable action
tags with immutable commit SHAs for each action usage (e.g., replace
docker/build-push-action@v7 and the other `@v6/`@v4/@v3 occurrences with their
corresponding full commit SHAs) so every "uses:" entry is pinned; update the 12
action instances referenced (including docker/build-push-action and the other
actions in the file) to their exact SHA values. Also add a top-level
permissions: block to the workflow that scopes the GITHUB_TOKEN to the minimal
required permissions for this workflow (declare only the specific permission
keys needed, e.g., read/write for specific resources used), ensuring the token
is not granted broad defaults. Ensure changes touch the workflow root
(top-level) and every "uses:" line that currently has a version tag.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
docker buildwithdocker/build-push-action@v7using GHA cache. Reads layers from the components-build-deploy workflow's cache scopes (frontend-amd64, etc.) so E2E gets warm layers from the last main build. Expected savings: 3-5 min.actions/cache@v4. Pin version in env var for cache key stability. ~15s saved per run.--build-tags=test(superset). ~30s saved.pip install junit2htmlwithpipx run junit2html(pre-installed on GHA runners). ~10s saved.Current PR wall-clock P50: ~10.4m (E2E bottleneck)
Expected PR wall-clock P50: ~5-7m
Test plan
:latestfor unchanged components🤖 Generated with Claude Code
Summary by CodeRabbit
Release Notes
Chores
Documentation