Skip to content

ci(site): drop Cloudflare purge, unblocking the deploy workflow #1114

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
xiaolai wants to merge 1 commit into
base: main
Choose a base branch
from
Open

ci(site): drop Cloudflare purge, unblocking the deploy workflow #1114

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
38 changes: 7 additions & 31 deletions .github/workflows/deploy-site.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,13 @@ name: Deploy nlpm.com
# - daily 23:00 UTC, after auditor-render-dashboard at 22:30 — so the
# dashboard's freshly-rendered cross-repo aggregate makes it onto
# the site without manual intervention.
#
# No Cloudflare purge step: nlpm.com is served directly by GitHub Pages
# on a DNS-only record, with no CF proxy in front. Proxying it broke
# GitHub's Let's Encrypt renewal — CF intercepted the ACME HTTP-01
# challenge, the cert expired 2026-08-17, and CF Full (Strict) then
# returned 526 to every visitor. Do not re-add the proxy or a purge
# step without re-checking that interaction first.

on:
push:
Expand Down Expand Up @@ -127,34 +134,3 @@ jobs:
# Deploys are inferred from gh-pages commits; we deliberately do
# not append to auditor/logs/events.jsonl on main, since that
# would push deploy churn into the auditor's append-only log.

- name: Wait for GitHub Pages CDN to propagate
# Pushing to gh-pages doesn't instantly update what GitHub Pages
# serves at its edge — there's typically a 30–60s lag. Purging
# CF before that lag elapses just re-caches the OLD content from
# GitHub's edge. Wait first, then purge.
run: sleep 75

- name: Purge Cloudflare cache
# Two failure modes a deploy can leave behind without this:
# - stale negative 404s for URLs that didn't exist on the prior
# deploy (CF caches 404s up to ~10 min)
# - stale content for URLs that existed but changed
# Both clear with a zone-wide purge_everything. Step gated on the
# three secrets so the workflow stays functional on forks.
env:
CF_API_KEY: ${{ secrets.CLOUDFLARE_API_KEY }}
CF_EMAIL: ${{ secrets.CLOUDFLARE_EMAIL }}
CF_ZONE_ID: ${{ secrets.CLOUDFLARE_ZONE_ID }}
run: |
if [ -z "$CF_API_KEY" ] || [ -z "$CF_EMAIL" ] || [ -z "$CF_ZONE_ID" ]; then
echo "Cloudflare secrets not configured — skipping purge."
exit 0
fi
RESP=$(curl -fsS -X POST \
-H "X-Auth-Key: $CF_API_KEY" \
-H "X-Auth-Email: $CF_EMAIL" \
-H "Content-Type: application/json" \
"https://api.cloudflare.com/client/v4/zones/$CF_ZONE_ID/purge_cache" \
--data '{"purge_everything":true}')
echo "$RESP" | python3 -c "import sys,json; d=json.load(sys.stdin); print('cf-purge success=' + str(d.get('success')))"
Loading