mirror of
https://github.com/nexu-io/open-design.git
synced 2026-05-31 19:04:39 +07:00
ci(landing): split landing deploy into staging gate + manual production (#2994)
* ci(landing): split landing deploy into staging gate + manual production A merge to `main` previously published the landing page straight to production (open-design.ai) via `landing-page-deploy`. There was no buffer to review the rendered site, so a bad merge was live instantly. Split deploys across two Cloudflare Pages projects so production is only ever reached by an explicit human action: - `landing-page-staging` (push to main) -> staging project `open-design-landing-staging` -> staging.open-design.ai. - `landing-page-production` (manual workflow_dispatch only) -> production project `open-design-landing` -> open-design.ai. Only this workflow names the production project; gate it with required reviewers on the `production` GitHub environment. - `landing-page-ci` now also deploys a per-PR preview into the staging project (`--branch=pr-<n>`) for same-repo branches and comments the URL. Fork PRs (no secrets / read-only token) skip the deploy and keep just the build validation. Path filters already scope this to landing edits. Decouple search-engine indexing from staging: - `blog-indexing-on-deploy` now triggers on `landing-page-production` (not every main push), so the test environment is never submitted to Google/IndexNow. - It diffs from a new `blog-indexed-prod` tag (the last indexed prod commit) instead of `HEAD^`, and force-advances the tag after a successful run, so a manual promotion bundling several merged posts indexes all of them rather than only the last commit. Staging and PR-preview builds drop `PUBLIC_GA_MEASUREMENT_ID` so test traffic does not pollute the production GA property. * ci(landing): keep staging + PR previews out of the search index staging.open-design.ai mirrors production and is exposed via cert transparency logs, so search engines can discover it. Indexing the mirror competes with open-design.ai for the same content. Emit `<meta name="robots" content="noindex, nofollow">` whenever OD_LANDING_NOINDEX=1, and set that flag on the staging and PR-preview builds (production leaves it unset and stays indexable). noindex is used rather than a robots.txt Disallow so crawlers can still fetch the page and read both the tag and the canonical, which already points at the production origin. * fix(landing): make staging noindex actually take effect The previous commit read `process.env.OD_LANDING_NOINDEX` directly in `seo-head.astro`, but `.astro` frontmatter is transformed by Vite and does not see process.env, so the meta never rendered. Two fixes: - Inject the flag as the compile-time constant `__OD_LANDING_NOINDEX__` via `vite.define` in astro.config.ts (config runs in Node and can read process.env); SeoHead consumes that constant. - The homepage (`index.astro`) and `og.astro` build their own <head> and never use SeoHead, so a per-component meta can miss pages. Add an `astro:build:done` integration that appends a catch-all `/* X-Robots-Tag: noindex, nofollow` to the Cloudflare Pages `_headers` on staging/preview builds, covering every response (homepage, assets, any custom-head page) at the HTTP layer. Production builds leave `_headers` untouched. Verified: build with OD_LANDING_NOINDEX=1 emits the _headers block and the SeoHead <meta>; build without the flag emits neither; astro check clean. * fix(landing): address review — pin prod checkout to main, defer index pointer Two blockers from review: - landing-page-production: workflow_dispatch can be launched from any ref via the Actions "Use workflow from" dropdown, so an operator could ship an arbitrary branch to open-design.ai. Pin the checkout to `ref: main` so the deployed artifact always equals reviewed main. - blog-indexing-on-deploy: the `blog-indexed-prod` pointer was advanced right after sitemap submission, before Inspect / Search Analytics / Render status / Open status PR. A failure in any of those still moved the pointer, so the next production run skipped those posts. Move the advance to the very end, gated on `success()`, so a failure leaves the tag in place and the range is re-processed next run (submissions are idempotent). * fix(landing): gate production promotion to the main ref only Follow-up to the production-path review note: pinning checkout to main fixed the deployed content, but the workflow was still dispatchable from any ref, which records a non-main production run and would dodge blog-indexing's `workflow_run` `branches: [main]` filter. Gate the whole job on `github.ref == 'refs/heads/main'` so a dispatch from any other branch/tag is skipped outright.
This commit is contained in:
lefarcen
GitHub
parent
6618780812
commit
7312c64580
11 changed files with 404 additions and 30 deletions
43
.github/workflows/blog-indexing-on-deploy.yml
vendored
43
.github/workflows/blog-indexing-on-deploy.yml
vendored
|
|
@ -1,6 +1,8 @@
|
|||
name: blog-indexing-on-deploy
|
||||
|
||||
# Runs after every successful `landing-page-deploy` on main. The job is
|
||||
# Runs after every successful `landing-page-production` promotion. Staging
|
||||
# deploys (`landing-page-staging`) intentionally do NOT trigger indexing, so
|
||||
# the test environment is never submitted to search engines. The job is
|
||||
# idempotent and follows the blog-indexing-automation skill:
|
||||
#
|
||||
# 1. Detect blog URLs added/modified in the deploy
|
||||
|
|
@ -16,7 +18,7 @@ name: blog-indexing-on-deploy
|
|||
|
||||
on:
|
||||
workflow_run:
|
||||
workflows: ['landing-page-deploy']
|
||||
workflows: ['landing-page-production']
|
||||
types: [completed]
|
||||
branches: [main]
|
||||
workflow_dispatch:
|
||||
|
|
@ -29,7 +31,10 @@ on:
|
|||
required: false
|
||||
|
||||
permissions:
|
||||
contents: read
|
||||
# `contents: write` lets the job advance the `blog-indexed-prod` tag after a
|
||||
# successful run so the next promotion diffs from exactly where this one
|
||||
# stopped. The status PR uses a separate GitHub App token (below).
|
||||
contents: write
|
||||
|
||||
concurrency:
|
||||
group: blog-indexing-on-deploy
|
||||
|
|
@ -147,8 +152,21 @@ jobs:
|
|||
mkdir -p .blog-indexing
|
||||
BASE="${{ github.event.inputs.base_sha || '' }}"
|
||||
if [ -z "$BASE" ]; then
|
||||
BASE="$(git rev-parse HEAD^)"
|
||||
# Diff from the last production deploy we already indexed, tracked
|
||||
# by the `blog-indexed-prod` tag. Production is a manual promotion
|
||||
# that may bundle several merged posts, so a HEAD^ diff would miss
|
||||
# all but the last commit. The tag is advanced at the end of a
|
||||
# successful run (see "Advance indexed-production pointer").
|
||||
git fetch --no-tags origin "+refs/tags/blog-indexed-prod:refs/tags/blog-indexed-prod" 2>/dev/null || true
|
||||
BASE="$(git rev-parse --verify --quiet 'refs/tags/blog-indexed-prod^{commit}' || true)"
|
||||
if [ -n "$BASE" ]; then
|
||||
echo "Diffing from last indexed production commit (blog-indexed-prod): $BASE"
|
||||
else
|
||||
BASE="$(git rev-parse HEAD^)"
|
||||
echo "No blog-indexed-prod tag yet; bootstrapping from HEAD^: $BASE"
|
||||
fi
|
||||
fi
|
||||
echo "base=$BASE" >> "$GITHUB_OUTPUT"
|
||||
pnpm --filter @open-design/landing-page exec tsx scripts/blog-indexing/detect-changed-urls.ts \
|
||||
--base "$BASE" \
|
||||
--head HEAD \
|
||||
|
|
@ -264,3 +282,20 @@ jobs:
|
|||
Generated by `.github/workflows/blog-indexing-on-deploy.yml`. The
|
||||
sidecar `docs/blog-indexing-status.json` is the canonical state;
|
||||
the markdown file is rendered from it.
|
||||
|
||||
# Advance the pointer LAST, only after every post-deploy step above
|
||||
# (detect, verify, submit, inspect, analytics, render, status PR)
|
||||
# succeeded. `success()` is false if any earlier step failed, so a
|
||||
# failure leaves the tag where it was and the next production run
|
||||
# re-processes the same range rather than silently skipping posts.
|
||||
# Skipped steps (count == 0, or GSC/bot not configured) do not count as
|
||||
# failures, so the pointer still advances over an empty/partial range.
|
||||
# Restricted to the real production-deploy trigger; an ad-hoc manual
|
||||
# dispatch must not move the production baseline.
|
||||
- name: Advance indexed-production pointer
|
||||
if: success() && github.event_name == 'workflow_run'
|
||||
run: |
|
||||
HEAD_SHA="$(git rev-parse HEAD)"
|
||||
git tag -f blog-indexed-prod "$HEAD_SHA"
|
||||
git push -f origin "refs/tags/blog-indexed-prod"
|
||||
echo "Advanced blog-indexed-prod -> $HEAD_SHA"
|
||||
|
|
|
|||
2
.github/workflows/ci.yml
vendored
2
.github/workflows/ci.yml
vendored
|
|
@ -87,7 +87,7 @@ jobs:
|
|||
case "$file" in
|
||||
*.md|*.mdx|*.txt|LICENSE|.gitignore|.editorconfig|.vscode/*|.idea/*|docs/*|.github/ISSUE_TEMPLATE/*|.github/CODEOWNERS)
|
||||
;;
|
||||
apps/landing-page/*|flake.nix|flake.lock|nix/*|.github/workflows/nix-check.yml|.github/workflows/landing-page-ci.yml|.github/workflows/landing-page-deploy.yml|.github/workflows/blog-indexing-on-deploy.yml|.github/workflows/blog-indexing-monitor.yml|.github/workflows/blog-3day-report.yml|.github/workflows/seo-daily-report.yml|.github/workflows/actionlint.yml|.github/workflows/visual-pr-capture.yml|.github/workflows/visual-pr-comment.yml)
|
||||
apps/landing-page/*|flake.nix|flake.lock|nix/*|.github/workflows/nix-check.yml|.github/workflows/landing-page-ci.yml|.github/workflows/landing-page-staging.yml|.github/workflows/landing-page-production.yml|.github/workflows/blog-indexing-on-deploy.yml|.github/workflows/blog-indexing-monitor.yml|.github/workflows/blog-3day-report.yml|.github/workflows/seo-daily-report.yml|.github/workflows/actionlint.yml|.github/workflows/visual-pr-capture.yml|.github/workflows/visual-pr-comment.yml)
|
||||
;;
|
||||
*)
|
||||
workspace_validation_required=true
|
||||
|
|
|
|||
87
.github/workflows/landing-page-ci.yml
vendored
87
.github/workflows/landing-page-ci.yml
vendored
|
|
@ -5,7 +5,8 @@ on:
|
|||
paths:
|
||||
# Workflow files
|
||||
- .github/workflows/landing-page-ci.yml
|
||||
- .github/workflows/landing-page-deploy.yml
|
||||
- .github/workflows/landing-page-staging.yml
|
||||
- .github/workflows/landing-page-production.yml
|
||||
- .github/workflows/blog-indexing-on-deploy.yml
|
||||
- .github/workflows/blog-indexing-monitor.yml
|
||||
# Landing page sources
|
||||
|
|
@ -28,7 +29,8 @@ on:
|
|||
- main
|
||||
paths:
|
||||
- .github/workflows/landing-page-ci.yml
|
||||
- .github/workflows/landing-page-deploy.yml
|
||||
- .github/workflows/landing-page-staging.yml
|
||||
- .github/workflows/landing-page-production.yml
|
||||
- .github/workflows/blog-indexing-on-deploy.yml
|
||||
- .github/workflows/blog-indexing-monitor.yml
|
||||
- apps/landing-page/**
|
||||
|
|
@ -44,6 +46,10 @@ on:
|
|||
|
||||
permissions:
|
||||
contents: read
|
||||
# Needed to post/update the preview-URL comment on the PR. Fork PRs run
|
||||
# with a read-only token regardless, so the preview steps below are gated
|
||||
# to same-repo branches.
|
||||
pull-requests: write
|
||||
|
||||
concurrency:
|
||||
group: landing-page-ci-${{ github.event.pull_request.number || github.ref }}
|
||||
|
|
@ -94,9 +100,13 @@ jobs:
|
|||
- name: Generate skill + template previews
|
||||
run: pnpm --filter @open-design/landing-page previews
|
||||
|
||||
# No PUBLIC_GA_MEASUREMENT_ID for PR/CI builds: the per-PR preview must
|
||||
# not report into the production GA property. OD_LANDING_NOINDEX=1 keeps
|
||||
# the PR preview (pr-<n>.open-design-landing-staging.pages.dev) out of
|
||||
# search engines. Only `landing-page-production` builds without these.
|
||||
- name: Build landing page
|
||||
env:
|
||||
PUBLIC_GA_MEASUREMENT_ID: ${{ vars.PUBLIC_GA_MEASUREMENT_ID }}
|
||||
OD_LANDING_NOINDEX: '1'
|
||||
run: pnpm --filter @open-design/landing-page build:static
|
||||
|
||||
- name: Lint changed blog SEO
|
||||
|
|
@ -153,3 +163,74 @@ jobs:
|
|||
process.exit(1);
|
||||
}
|
||||
NODE
|
||||
|
||||
# --- PR preview deploy -------------------------------------------------
|
||||
# Publish this PR's built site to its own preview URL in the STAGING
|
||||
# project (`--branch=pr-<number>`) so reviewers see the rendered result
|
||||
# before merge. It lands in the staging project, never the production
|
||||
# project. Gated to same-repo branches: fork PRs run without the
|
||||
# Cloudflare secrets and with a read-only token, so they skip the
|
||||
# deploy/comment and keep just the validation above.
|
||||
- name: Deploy PR preview to Cloudflare Pages
|
||||
id: preview
|
||||
if: github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name == github.repository
|
||||
uses: cloudflare/wrangler-action@v3
|
||||
with:
|
||||
apiToken: ${{ secrets.CLOUDFLARE_API_TOKEN }}
|
||||
accountId: ${{ secrets.CLOUDFLARE_ACCOUNT_ID }}
|
||||
workingDirectory: apps/landing-page
|
||||
packageManager: npm
|
||||
command: >
|
||||
pages deploy out
|
||||
--project-name=open-design-landing-staging
|
||||
--branch=pr-${{ github.event.pull_request.number }}
|
||||
|
||||
- name: Comment preview URL on PR
|
||||
if: github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name == github.repository
|
||||
uses: actions/github-script@v7
|
||||
env:
|
||||
DEPLOY_URL: ${{ steps.preview.outputs.deployment-url }}
|
||||
ALIAS_URL: ${{ steps.preview.outputs.pages-deployment-alias-url }}
|
||||
with:
|
||||
script: |
|
||||
const marker = '<!-- landing-preview -->';
|
||||
const deploy = process.env.DEPLOY_URL || '';
|
||||
const alias =
|
||||
process.env.ALIAS_URL ||
|
||||
`https://pr-${context.issue.number}.open-design-landing-staging.pages.dev`;
|
||||
const sha = context.payload.pull_request.head.sha.slice(0, 7);
|
||||
const body = [
|
||||
marker,
|
||||
'### 🚀 Landing page preview',
|
||||
'',
|
||||
'This PR is deployed to a Cloudflare Pages preview — **not** staging or production:',
|
||||
'',
|
||||
`- Stable alias: ${alias}`,
|
||||
deploy ? `- This build: ${deploy}` : '',
|
||||
'',
|
||||
`Updated for commit \`${sha}\`.`,
|
||||
]
|
||||
.filter(Boolean)
|
||||
.join('\n');
|
||||
const comments = await github.paginate(github.rest.issues.listComments, {
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: context.issue.number,
|
||||
per_page: 100,
|
||||
});
|
||||
const existing = comments.find((c) => c.body && c.body.includes(marker));
|
||||
if (existing) {
|
||||
await github.rest.issues.updateComment({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
comment_id: existing.id,
|
||||
body,
|
||||
});
|
||||
} else {
|
||||
await github.rest.issues.createComment({
|
||||
owner: context.repo.owner,
|
||||
repo: context.repo.repo,
|
||||
issue_number: context.issue.number,
|
||||
body,
|
||||
});
|
||||
}
|
||||
|
|
|
|||
156
.github/workflows/landing-page-production.yml
vendored
Normal file
156
.github/workflows/landing-page-production.yml
vendored
Normal file
|
|
@ -0,0 +1,156 @@
|
|||
name: landing-page-production
|
||||
|
||||
# Promotes the current landing page to PRODUCTION: the `open-design-landing`
|
||||
# Cloudflare Pages project, served at open-design.ai. This is the ONLY
|
||||
# workflow that names the production project, and it is manual-only
|
||||
# (workflow_dispatch) — a merge to `main` can never reach production on its
|
||||
# own; it only updates the staging project (staging.open-design.ai) via
|
||||
# `landing-page-staging`. Gate this further by configuring required reviewers
|
||||
# on the GitHub `production` environment (Settings → Environments).
|
||||
#
|
||||
# The build is identical to staging/CI, so what you reviewed on
|
||||
# staging.open-design.ai is what ships.
|
||||
|
||||
on:
|
||||
workflow_dispatch:
|
||||
inputs:
|
||||
reason:
|
||||
description: 'Why promote now? (recorded in the run log)'
|
||||
required: false
|
||||
|
||||
permissions:
|
||||
contents: read
|
||||
deployments: write
|
||||
|
||||
# Never cancel an in-flight production deploy.
|
||||
concurrency:
|
||||
group: landing-page-production
|
||||
cancel-in-progress: false
|
||||
|
||||
jobs:
|
||||
deploy:
|
||||
name: Deploy landing page to production
|
||||
# Production ships `main` only. workflow_dispatch can be launched from any
|
||||
# ref via the Actions "Use workflow from" dropdown; gate the whole job on
|
||||
# the main ref so a dispatch from a feature branch/tag is skipped outright
|
||||
# (no deploy) instead of recording a non-main production run — which would
|
||||
# also dodge blog-indexing's `workflow_run` `branches: [main]` filter.
|
||||
if: github.repository == 'nexu-io/open-design' && github.ref == 'refs/heads/main'
|
||||
runs-on: ubuntu-latest
|
||||
timeout-minutes: 20
|
||||
environment:
|
||||
name: production
|
||||
url: https://open-design.ai
|
||||
|
||||
steps:
|
||||
- name: Checkout
|
||||
uses: actions/checkout@v6.0.2
|
||||
with:
|
||||
# Production always ships `main`. workflow_dispatch can be launched
|
||||
# from any ref via the Actions "Use workflow from" dropdown, so pin
|
||||
# the checkout to main — the deployed artifact must equal reviewed
|
||||
# main, never whatever branch/tag the operator happened to select.
|
||||
ref: main
|
||||
|
||||
- name: Setup pnpm
|
||||
uses: pnpm/action-setup@v5
|
||||
with:
|
||||
version: 10.33.2
|
||||
|
||||
- name: Setup Node.js
|
||||
uses: actions/setup-node@v6
|
||||
with:
|
||||
node-version: 24
|
||||
cache: pnpm
|
||||
|
||||
- name: Install dependencies
|
||||
run: pnpm install --frozen-lockfile
|
||||
|
||||
- name: Resolve Playwright version
|
||||
id: playwright-version
|
||||
run: |
|
||||
version=$(node -p "require('./apps/landing-page/package.json').devDependencies.playwright.replace(/[^0-9.]/g,'')")
|
||||
echo "version=$version" >> "$GITHUB_OUTPUT"
|
||||
|
||||
- name: Cache generated previews
|
||||
id: previews-cache
|
||||
uses: actions/cache@v5.0.5
|
||||
with:
|
||||
path: apps/landing-page/public/previews
|
||||
key: landing-page-previews-${{ runner.os }}-${{ hashFiles('pnpm-lock.yaml', 'package.json', 'apps/landing-page/package.json', 'apps/landing-page/scripts/generate-previews.ts', 'skills/**', 'design-templates/**', 'templates/live-artifacts/**') }}
|
||||
restore-keys: |
|
||||
landing-page-previews-${{ runner.os }}-
|
||||
|
||||
- name: Cache Playwright browsers
|
||||
uses: actions/cache@v5.0.5
|
||||
with:
|
||||
path: ~/.cache/ms-playwright
|
||||
key: playwright-${{ runner.os }}-${{ steps.playwright-version.outputs.version }}
|
||||
|
||||
- name: Install Playwright Chromium
|
||||
run: pnpm --filter @open-design/landing-page exec playwright install --with-deps chromium
|
||||
|
||||
- name: Typecheck landing page
|
||||
run: pnpm --filter @open-design/landing-page typecheck
|
||||
|
||||
# Generate previews before build so they end up in `out/previews/`.
|
||||
# Soft vs. hard failure is enforced inside the script itself:
|
||||
# individual broken `example.html` entries are logged and skipped,
|
||||
# but a systemic failure (chromium launch error, every job failing)
|
||||
# exits non-zero so we don't silently ship a deploy with zero
|
||||
# thumbnails to production.
|
||||
- name: Generate skill + template previews
|
||||
run: pnpm --filter @open-design/landing-page previews
|
||||
|
||||
- name: Build landing page
|
||||
env:
|
||||
PUBLIC_GA_MEASUREMENT_ID: ${{ vars.PUBLIC_GA_MEASUREMENT_ID }}
|
||||
run: pnpm --filter @open-design/landing-page build:static
|
||||
|
||||
- name: Verify zero external JavaScript
|
||||
run: |
|
||||
node <<'NODE'
|
||||
const { readFileSync } = require('node:fs');
|
||||
const html = readFileSync('apps/landing-page/out/index.html', 'utf8');
|
||||
const forbidden = [
|
||||
/<script\b[^>]*\bsrc=/i,
|
||||
/type=["']module["']/i,
|
||||
/\/_astro\/[^"'<>\s]+\.js/i,
|
||||
];
|
||||
for (const pattern of forbidden) {
|
||||
if (pattern.test(html)) {
|
||||
console.error(`Unexpected client JavaScript matched ${pattern}`);
|
||||
process.exit(1);
|
||||
}
|
||||
}
|
||||
NODE
|
||||
|
||||
- name: Verify Cloudflare image resizing URLs
|
||||
run: |
|
||||
node <<'NODE'
|
||||
const { readFileSync } = require('node:fs');
|
||||
const html = readFileSync('apps/landing-page/out/index.html', 'utf8');
|
||||
const resizedUrls = html.match(/https:\/\/static\.open-design\.ai\/cdn-cgi\/image\//g) ?? [];
|
||||
if (resizedUrls.length < 16) {
|
||||
console.error(`Expected at least 16 Cloudflare resized image URLs, found ${resizedUrls.length}`);
|
||||
process.exit(1);
|
||||
}
|
||||
if (/(?:src|content)=["']\/assets\/[A-Za-z0-9_.-]+\.png/.test(html)) {
|
||||
console.error('Found local /assets/*.png image reference in generated landing HTML.');
|
||||
process.exit(1);
|
||||
}
|
||||
NODE
|
||||
|
||||
# `--branch=main` IS the Cloudflare Pages production branch, so this
|
||||
# publishes to the production domain (open-design.ai).
|
||||
- name: Deploy to Cloudflare Pages (production)
|
||||
uses: cloudflare/wrangler-action@v3
|
||||
with:
|
||||
apiToken: ${{ secrets.CLOUDFLARE_API_TOKEN }}
|
||||
accountId: ${{ secrets.CLOUDFLARE_ACCOUNT_ID }}
|
||||
workingDirectory: apps/landing-page
|
||||
packageManager: npm
|
||||
command: >
|
||||
pages deploy out
|
||||
--project-name=open-design-landing
|
||||
--branch=main
|
||||
|
|
@ -1,4 +1,11 @@
|
|||
name: landing-page-deploy
|
||||
name: landing-page-staging
|
||||
|
||||
# Pushes to `main` deploy the landing page to the STAGING Cloudflare Pages
|
||||
# project (`open-design-landing-staging`, served at staging.open-design.ai).
|
||||
# Production (the separate `open-design-landing` project → open-design.ai) is
|
||||
# never named here — it is reached exclusively through the manual
|
||||
# `landing-page-production` workflow. This is the safety gate: project
|
||||
# separation means a merge to main can only ever change staging.
|
||||
|
||||
on:
|
||||
push:
|
||||
|
|
@ -6,7 +13,8 @@ on:
|
|||
- main
|
||||
paths:
|
||||
# Workflow files
|
||||
- .github/workflows/landing-page-deploy.yml
|
||||
- .github/workflows/landing-page-staging.yml
|
||||
- .github/workflows/landing-page-production.yml
|
||||
- .github/workflows/landing-page-ci.yml
|
||||
# Landing page sources
|
||||
- apps/landing-page/**
|
||||
|
|
@ -31,15 +39,18 @@ permissions:
|
|||
deployments: write
|
||||
|
||||
concurrency:
|
||||
group: landing-page-deploy-${{ github.ref }}
|
||||
group: landing-page-staging-${{ github.ref }}
|
||||
cancel-in-progress: true
|
||||
|
||||
jobs:
|
||||
deploy:
|
||||
name: Deploy landing page
|
||||
name: Deploy landing page to staging
|
||||
if: github.repository == 'nexu-io/open-design'
|
||||
runs-on: ubuntu-latest
|
||||
timeout-minutes: 20
|
||||
environment:
|
||||
name: landing-staging
|
||||
url: https://staging.open-design.ai
|
||||
|
||||
steps:
|
||||
- name: Checkout
|
||||
|
|
@ -95,9 +106,15 @@ jobs:
|
|||
- name: Generate skill + template previews
|
||||
run: pnpm --filter @open-design/landing-page previews
|
||||
|
||||
# No PUBLIC_GA_MEASUREMENT_ID on staging: leaving it unset omits the
|
||||
# Google Analytics tag entirely, so test-environment traffic does not
|
||||
# pollute the production GA property. OD_LANDING_NOINDEX=1 makes every
|
||||
# page emit <meta name="robots" content="noindex"> so the staging mirror
|
||||
# stays out of search engines. Only `landing-page-production` builds
|
||||
# without these (GA on, indexable).
|
||||
- name: Build landing page
|
||||
env:
|
||||
PUBLIC_GA_MEASUREMENT_ID: ${{ vars.PUBLIC_GA_MEASUREMENT_ID }}
|
||||
OD_LANDING_NOINDEX: '1'
|
||||
run: pnpm --filter @open-design/landing-page build:static
|
||||
|
||||
- name: Verify zero external JavaScript
|
||||
|
|
@ -134,7 +151,11 @@ jobs:
|
|||
}
|
||||
NODE
|
||||
|
||||
- name: Deploy to Cloudflare Pages
|
||||
# Deploys to the dedicated STAGING project. `--branch=main` is that
|
||||
# project's production branch, so it serves at its custom domain
|
||||
# staging.open-design.ai. The production project (open-design-landing)
|
||||
# is a different project and is never named here.
|
||||
- name: Deploy to Cloudflare Pages (staging)
|
||||
uses: cloudflare/wrangler-action@v3
|
||||
with:
|
||||
apiToken: ${{ secrets.CLOUDFLARE_API_TOKEN }}
|
||||
|
|
@ -143,5 +164,5 @@ jobs:
|
|||
packageManager: npm
|
||||
command: >
|
||||
pages deploy out
|
||||
--project-name=open-design-landing
|
||||
--branch=${{ github.ref_name }}
|
||||
--project-name=open-design-landing-staging
|
||||
--branch=main
|
||||
|
|
@ -90,10 +90,30 @@ Tightly coupled with:
|
|||
upstream Markdown (e.g., `guizang-ppt`) doesn't break the build
|
||||
when an author uses a slightly different `od:` key.
|
||||
|
||||
## Auto-deploy contract
|
||||
## Deploy contract (staging → manual production)
|
||||
|
||||
`.github/workflows/landing-page-deploy.yml` runs on push to `main`
|
||||
when **any** of these change:
|
||||
Deploys are split across **two Cloudflare Pages projects** so a merge to
|
||||
`main` can never publish to the live site on its own:
|
||||
|
||||
- Production project `open-design-landing` → `open-design.ai`.
|
||||
- Staging project `open-design-landing-staging` → `staging.open-design.ai`.
|
||||
|
||||
The safety gate is project separation: only the manual production workflow
|
||||
ever names the production project.
|
||||
|
||||
- `.github/workflows/landing-page-staging.yml` runs on push to `main` and
|
||||
deploys to the **staging project** (`open-design-landing-staging`,
|
||||
`staging.open-design.ai`).
|
||||
- `.github/workflows/landing-page-production.yml` is **manual**
|
||||
(`workflow_dispatch`) and is the only workflow that names the production
|
||||
project (`open-design-landing`, `open-design.ai`). Gate it with required
|
||||
reviewers on the GitHub `production` environment.
|
||||
- `.github/workflows/landing-page-ci.yml` runs on PRs: it validates the build
|
||||
and, for same-repo branches, deploys a per-PR preview into the staging
|
||||
project (`--branch=pr-<number>` →
|
||||
`pr-<number>.open-design-landing-staging.pages.dev`) and comments the URL.
|
||||
|
||||
The staging workflow triggers when **any** of these change:
|
||||
|
||||
- `apps/landing-page/**`
|
||||
- `design-templates/open-design-landing/**`
|
||||
|
|
@ -102,12 +122,12 @@ when **any** of these change:
|
|||
- `craft/**`
|
||||
- `templates/**`
|
||||
- `package.json`, `pnpm-lock.yaml`, `pnpm-workspace.yaml`
|
||||
- the workflow file itself
|
||||
- the workflow files themselves
|
||||
|
||||
A push that only edits a SKILL.md MUST trigger this workflow — if it
|
||||
doesn't, the `paths:` filter has drifted from the content-collection
|
||||
glob and the published site will fall behind silently. Treat that as
|
||||
a regression, not a feature.
|
||||
A push that only edits a SKILL.md MUST trigger the staging workflow — if it
|
||||
doesn't, the `paths:` filter has drifted from the content-collection glob and
|
||||
the staged site will fall behind silently. Treat that as a regression, not a
|
||||
feature.
|
||||
|
||||
## Common commands
|
||||
|
||||
|
|
|
|||
|
|
@ -55,6 +55,18 @@ export interface SeoHeadProps {
|
|||
|
||||
const props = Astro.props as SeoHeadProps;
|
||||
|
||||
// Staging / PR-preview builds set OD_LANDING_NOINDEX=1 so the mirror at
|
||||
// staging.open-design.ai (exposed via certificate-transparency logs) is kept
|
||||
// out of the search index. We emit `noindex` rather than a robots.txt
|
||||
// `Disallow` so crawlers can still fetch the page and read both this tag and
|
||||
// the canonical (which points at the production origin). Production builds
|
||||
// leave the flag unset and stay fully indexable.
|
||||
//
|
||||
// `__OD_LANDING_NOINDEX__` is a compile-time constant injected by
|
||||
// `vite.define` in astro.config.ts — `.astro` frontmatter is transformed by
|
||||
// Vite and cannot read process.env directly.
|
||||
const noindex = __OD_LANDING_NOINDEX__;
|
||||
|
||||
const SITE_NAME = 'Open Design';
|
||||
const TAGLINE = 'Design with the agent already on your laptop.';
|
||||
const isArticle = props.kind === 'article';
|
||||
|
|
@ -149,6 +161,7 @@ const blogJsonLd =
|
|||
<title>{fullTitle}</title>
|
||||
<meta name='description' content={props.description} />
|
||||
<meta name='theme-color' content='#efe7d2' />
|
||||
{noindex && <meta name='robots' content='noindex, nofollow' />}
|
||||
<link rel='canonical' href={canonical} />
|
||||
{alternateLinks.map((entry) => (
|
||||
<link rel='alternate' hreflang={entry.hreflang} href={entry.href} />
|
||||
|
|
|
|||
4
apps/landing-page/app/env.d.ts
vendored
4
apps/landing-page/app/env.d.ts
vendored
|
|
@ -1 +1,5 @@
|
|||
/// <reference path="../.astro/types.d.ts" />
|
||||
|
||||
// Compile-time constant injected by `vite.define` in astro.config.ts. True on
|
||||
// staging / PR-preview builds (OD_LANDING_NOINDEX=1), false in production.
|
||||
declare const __OD_LANDING_NOINDEX__: boolean;
|
||||
|
|
|
|||
|
|
@ -1,8 +1,9 @@
|
|||
import sitemap, { type SitemapItem } from '@astrojs/sitemap';
|
||||
import { readFileSync, readdirSync } from 'node:fs';
|
||||
import { appendFileSync, readFileSync, readdirSync } from 'node:fs';
|
||||
import { join } from 'node:path';
|
||||
import type { AstroUserConfig } from 'astro';
|
||||
import { defineConfig } from 'astro/config';
|
||||
import type { AstroIntegration } from 'astro';
|
||||
import {
|
||||
DEFAULT_LOCALE,
|
||||
LANDING_LOCALES,
|
||||
|
|
@ -109,6 +110,34 @@ const editorialPaperTheme: ShikiThemeObject = {
|
|||
// builds (Cloudflare Pages preview deployments, local previews on a
|
||||
// different host) can stamp their own URL without forking the config.
|
||||
const site = process.env.OD_LANDING_SITE ?? 'https://open-design.ai';
|
||||
// Staging / PR-preview builds set OD_LANDING_NOINDEX=1. Resolved here (config
|
||||
// runs in Node and can read process.env) and inlined into components as the
|
||||
// compile-time constant `__OD_LANDING_NOINDEX__` via vite.define below —
|
||||
// `.astro` frontmatter is transformed by Vite and cannot read process.env.
|
||||
const landingNoindex = process.env.OD_LANDING_NOINDEX === '1';
|
||||
|
||||
// Staging / PR-preview only: append a catch-all `X-Robots-Tag: noindex` to the
|
||||
// Cloudflare Pages `_headers` so EVERY response stays out of search indexes —
|
||||
// including the React-rendered homepage and `og.astro`, which build their own
|
||||
// <head> and don't go through SeoHead (whose <meta robots> covers HTML pages).
|
||||
// Production builds (flag unset) leave `_headers` untouched.
|
||||
const noindexHeaders: AstroIntegration = {
|
||||
name: 'staging-noindex-headers',
|
||||
hooks: {
|
||||
'astro:build:done': () => {
|
||||
if (!landingNoindex) return;
|
||||
// `out/_headers` already exists (copied verbatim from public/_headers
|
||||
// during the build). Append a catch-all noindex header. Path is built
|
||||
// from the config dir + outDir rather than the hook's `dir` URL to
|
||||
// avoid any URL-resolution ambiguity.
|
||||
appendFileSync(
|
||||
join(import.meta.dirname, 'out', '_headers'),
|
||||
'\n# Staging / preview mirror — keep out of search indexes.\n/*\n X-Robots-Tag: noindex, nofollow\n',
|
||||
);
|
||||
},
|
||||
},
|
||||
};
|
||||
|
||||
const sitemapLocales = Object.fromEntries(
|
||||
LANDING_LOCALES.map((locale) => [locale.code, locale.htmlLang]),
|
||||
);
|
||||
|
|
@ -137,6 +166,11 @@ export default defineConfig({
|
|||
srcDir: './app',
|
||||
outDir: './out',
|
||||
trailingSlash: 'always',
|
||||
vite: {
|
||||
define: {
|
||||
__OD_LANDING_NOINDEX__: JSON.stringify(landingNoindex),
|
||||
},
|
||||
},
|
||||
build: {
|
||||
// Inline every emitted stylesheet directly into the HTML <head>.
|
||||
// Trade-off: HTML pages grow by ~10-15KB (already Brotli-compressed
|
||||
|
|
@ -159,6 +193,7 @@ export default defineConfig({
|
|||
},
|
||||
},
|
||||
integrations: [
|
||||
noindexHeaders,
|
||||
sitemap({
|
||||
i18n: {
|
||||
defaultLocale: DEFAULT_LOCALE,
|
||||
|
|
|
|||
|
|
@ -8,7 +8,7 @@
|
|||
* - page is present in sitemap output
|
||||
*
|
||||
* Polls each URL with a short backoff so the script can be invoked
|
||||
* immediately after `landing-page-deploy` completes (Cloudflare Pages
|
||||
* immediately after `landing-page-production` completes (Cloudflare Pages
|
||||
* propagation usually < 60 s but not guaranteed).
|
||||
*
|
||||
* Usage: tsx verify-readiness.ts --urls <file.json> [--out file.json] [--timeout-ms 180000]
|
||||
|
|
|
|||
|
|
@ -14,7 +14,7 @@ doc is its concrete implementation in `nexu-io/open-design`.
|
|||
| Trigger | Job | Outcome |
|
||||
|---|---|---|
|
||||
| `landing-page-ci` | `lint-blog-seo.ts` + `check-blog-url-changes.ts` | Changed posts are checked for frontmatter, internal/external links, rendered canonical/JSON-LD/OG metadata, and slug delete/rename redirects before they can merge. |
|
||||
| `landing-page-deploy` finishes successfully on `main` | `blog-indexing-on-deploy.yml` | New blog URLs are detected, verified ready, submitted to IndexNow, the sitemap-index is re-submitted to GSC, baseline URL Inspection is captured, and baseline Search Analytics is queried. |
|
||||
| `landing-page-production` promotion finishes successfully | `blog-indexing-on-deploy.yml` | New blog URLs are detected, verified ready, submitted to IndexNow, the sitemap-index is re-submitted to GSC, baseline URL Inspection is captured, and baseline Search Analytics is queried. Staging deploys (`landing-page-staging`) never trigger this. |
|
||||
| Daily `cron: 0 2 * * *` | `blog-indexing-monitor.yml` | Every blog post in the T+1 / T+3 / T+7 / T+14 window is re-inspected; GSC Search Analytics is refreshed; stall and low-traffic issues are opened/refreshed when needed. |
|
||||
| Daily `cron: 0 2 * * *` (10:00 Asia/Shanghai) | `blog-3day-report.yml` | T-3 cohort + 30-day rolling cohort traffic digest written to `docs/blog-traffic-digest.md` via the `automation/blog-traffic-digest` PR, with an optional Feishu group push. Read-only against GSC. |
|
||||
| Manual `workflow_dispatch` | `blog-indexing-monitor.yml` | Maintainers can dry-run or explicitly publish a token-gated dev.to/Hashnode cross-post with canonical URL pointing back to Open Design. |
|
||||
|
|
@ -58,10 +58,19 @@ content / SEO issue.
|
|||
|
||||
## Architecture
|
||||
|
||||
Because production is a manual promotion that can bundle several merged
|
||||
posts, `detect-changed-urls` diffs from the `blog-indexed-prod` git tag (the
|
||||
last commit this workflow successfully indexed) rather than `HEAD^`. The tag
|
||||
is force-advanced to the deployed commit at the end of a successful run, so
|
||||
the next promotion picks up exactly the posts merged in between. On the very
|
||||
first run the tag does not exist yet and the workflow bootstraps from `HEAD^`
|
||||
— create the tag manually (`git tag blog-indexed-prod <sha>; git push origin
|
||||
blog-indexed-prod`) if an initial multi-commit backfill is needed.
|
||||
|
||||
```
|
||||
landing-page-deploy ──success──▶ blog-indexing-on-deploy
|
||||
landing-page-production ──success──▶ blog-indexing-on-deploy
|
||||
│
|
||||
detect-changed-urls
|
||||
detect-changed-urls (base = blog-indexed-prod tag)
|
||||
│
|
||||
verify-readiness (200 / canonical / sitemap)
|
||||
│
|
||||
|
|
@ -224,7 +233,7 @@ The expected steady state:
|
|||
- Renames are handled as both a redirect requirement for the old slug
|
||||
and a newly deployed URL for the destination slug, so the new page is
|
||||
included in the post-deploy readiness and baseline inspection flow.
|
||||
- New post ships → `landing-page-deploy` runs → `blog-indexing-on-deploy`
|
||||
- New post ships → `landing-page-production` promotion runs → `blog-indexing-on-deploy`
|
||||
runs → IndexNow is called, GSC sitemap is submitted, and the bot PR
|
||||
opens with the baseline verdict plus any available 7d/28d traffic
|
||||
metrics.
|
||||
|
|
|
|||
Loading…
Reference in a new issue