Day 44 - March 27, 2026

Daily Work Report (Mar 27, 2026)

Work Summary

Worked on Cloud Job Sub-App deployment stability and resolved a critical build blocker (ERR_PNPM_UNSUPPORTED_ENGINE) caused by Node.js engine mismatch. Aligned the Docker runtime with project engine requirements (Node 24), enabled corepack for consistent pnpm behavior, and validated backend-focused deployment flow for Railway. Also completed runtime stabilization by removing a non-essential failing OpenCV import path and confirming service health recovery after deployment retries.

Hours Worked

8.0

Show Your Work (References)

Dockerfile: Updated base image to node:24.14.0-slim; enabled corepack; kept frozen dependency install path.
package.json: Verified engines compatibility (node >=24.0.0) with cloud and container runtime.
railway.toml: Corrected build section placement so custom backend build steps are honored.
faces.py: Removed blocking OpenCV-related import path not required for active routing.
main.py and Railway deploy logs: Verified startup import path and final healthcheck recovery.

Learnings / Outcomes

Environment parity is critical: container Node version must satisfy engines constraints exactly.
corepack is a reliable way to manage pnpm versions in CI/CD and Docker without global installs.
Keeping deployment scope backend-only prevents monorepo frontend engine conflicts during Railway builds.
Native library requirements should be verified early for vision-capable services in cloud images.

Blockers / Risks

Blocker (resolved): Docker build failed due to Node.js v18 vs required v24.
Blocker (resolved): Runtime crash from OpenCV/native dependency path during startup.
Ongoing risk: Cloud environments drifting from pinned runtime versions can reintroduce engine and native-lib failures.
Mitigation: Keep pinned Docker base versions, strict engine checks, and explicit native dependency installation in deployment build steps.

Skills Used

Docker, Node.js v24, pnpm, Railway, Python 3.13 deployment, Docker/Nixpacks, OpenCV dependency debugging, Cloud Healthchecks, DevOps Troubleshooting

Next Step

Test live log streaming from the Hugging Face cloud container to validate end-to-end job observability.