ci: enable worker leg in production deploy (dual-target) — Phase 2 Part 3#1729
Conversation
Production release 14 july (2)
Production release (3)
Production release (4)
Uncomment the worker matrix leg so a push to main deploys to both the web and worker tiers in parallel. Worker reachable via PROD_SSH_WORKER_HOST (SSH, same Key A + apache-sudo path as web). Phase 2 Part 3. Ref: coloredcow-admin/goonj-crm#342
WalkthroughThe production deployment workflow now runs its SSH deployment step for both web and worker targets by activating the worker entry in the deployment matrix. ChangesProduction deployment
Estimated code review effort: 1 (Trivial) | ~2 minutes Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
🧹 Nitpick comments (1)
.github/workflows/production-deployment.yml (1)
18-18: 🗄️ Data Integrity & Integration | 🔵 TrivialKeep shared production mutations out of the matrix job.
Enabling the worker causes every matrix step to run twice in parallel. The shown Git sync is host-local, but verify that no subsequent steps perform database migrations, schema changes, cache flushes, or other singleton operations; those must run in a separate one-time job.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.github/workflows/production-deployment.yml at line 18, Update the production deployment workflow’s matrix configuration around the worker entry so host-local Git synchronization remains parallelizable, while any subsequent singleton operations such as migrations, schema changes, or cache flushes are removed from the matrix path and moved into a separate one-time deployment job. Preserve independent per-host deployment steps for the matrix targets.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In @.github/workflows/production-deployment.yml:
- Line 18: Update the production deployment workflow’s matrix configuration
around the worker entry so host-local Git synchronization remains
parallelizable, while any subsequent singleton operations such as migrations,
schema changes, or cache flushes are removed from the matrix path and moved into
a separate one-time deployment job. Preserve independent per-host deployment
steps for the matrix targets.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Repository UI
Review profile: CHILL
Plan: Pro
Run ID: 33e5b4ff-5d3e-4bcc-b4ae-828ac54d731b
📒 Files selected for processing (1)
.github/workflows/production-deployment.yml
Uncomments the worker matrix leg in
production-deployment.yml, so a push tomaindeploys to both the web and worker tiers in parallel (fail-fast: false, workflow-levelconcurrency).Worker target:
PROD_SSH_WORKER_HOST= Elastic IP13.207.97.62(stable). Reached over SSH — same mechanism as web (login viaPROD_SSH_PRIVATE_KEY→git reset --hardasapachevia sudo). Key A verified authenticating to the worker via the EIP; the worker inherits the same git checkout/docroot from the prod clone.Before merging
git reset --hardruns per host; any CiviCRM DB-upgrade/schema/cache step must run once, not per-host (see the deploy runbook).Draft until cutover timing.
Ref: coloredcow-admin/goonj-crm#342
Summary by CodeRabbit