Step 8: Production
Deploy production the same way as staging, behind the production environment's approval.
Goal: production running from the same commit that passed on staging.
Production is a separate set of resources with its own state file, passphrase, VPC, databases, bucket, Droplets, keys and secrets. Nothing is shared with staging except the Cloudflare zone stack and, if you chose so, the API tokens.
- Confirm the
productionGitHub environment has its secrets and variables (step 4), includingAPI_DNS_PROXIED=falsefor the first run. - Actions → Deploy → Run workflow → environment:
production, frommain. A required reviewer approves each job that uses the environment. - Do the checks in step 6.3 with the production hostnames.
- When the App Platform domain is Active, delete
API_DNS_PROXIEDand run the Deploy workflow for production again. - Create the owner's password and run the smoke test as in step 7 with production hostnames.
- Run the full production checklist and the security checklist before real traffic.
From now on, production changes only by an approved Deploy run. Next: Step 9: Operate.
Step 7: First sign-in
Create the owner's console password, sign in through Cloudflare Access, create a project and key, and run the control-plane smoke test.
Step 9: Operate
Change infrastructure by pull request, release and roll back, rotate generated secrets, run OpenTofu locally and import or remove resources.