feat(deploy): orchestrate deployment readiness, decouple backend liveness, and fix PrismaPg

[DerekRoberts]

Overview

This PR optimizes OpenShift template deployments, container orchestration, and health probing across the database, backend, and frontend services. It resolves a critical runtime schema resolution bug in the backend, decouples backend liveness checks from downstream database status to prevent cascading outages, replaces high-overhead shell-out probes with lightweight native checks, and blocks the frontend rollout until the backend service is ready.

There are no CI/workflow changes; the e2e test environment stability is achieved purely through corrected Kubernetes orchestration and templates.

Changes

Frontend (frontend/openshift.deploy.yml)

• Backend Syncing: Added a wait-for-backend initContainer to hold the frontend rollout until the backend's /api/health endpoint responds with success. This cleanly prevents the frontend from starting up when the backend is not yet ready to serve requests.
• Probe Bypass: Exposed port 3001 (health) and switched startup, readiness, and liveness probes to query /health on that port, bypassing the Coraza WAF middleware and avoiding log clutter.

Backend

• Database Schema Fix (backend/src/prisma.service.ts): Passed { schema: DB_SCHEMA } to the PrismaPg driver adapter. Previously, the custom Postgres schema was ignored because the pg Pool wrapper does not parse connection string query parameters, resulting in runtime 500 errors.
• Liveness Decoupling (backend/openshift.deploy.yml): Swapped the livenessProbe target from /api/health (database-dependent) to /api (in-memory greeting). The pod will no longer restart itself to death during transient database outages.
• Placeholder Cleanup (backend/src/app.service.ts, specs, and E2E): Replaced the generic placeholder greeting 'Hello Backend!' with 'Backend is live!'.

Database (common/openshift.database.yml)

• Lightweight Probing: Replaced the high-overhead psql -c 'SELECT 1' shell-out probes with native pg_isready execution, reducing resource consumption during routine checks.

---

Thanks for the PR!

Deployments, as required, will be available below:

• Frontend
• Backend

Please create PRs in draft mode. Mark as ready to enable:

• Analysis Workflow

After merge, new images are deployed in:

• Merge Workflow

[Copilot]

Pull request overview

Updates the reusable CI E2E workflow to wait for the dynamically deployed frontend route to become reachable/healthy before starting Playwright runs, reducing flakiness/timeouts (notably Chromium) caused by route propagation delays.

Changes:

• Add a “Wait for frontend health check” step that polls the PR frontend URL until it returns HTTP 200 (or times out).
• Run Playwright E2E tests only after the frontend endpoint is confirmed available.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.