pipeline-build-errors

Prompt file imported from drbf17/datathon (.github/prompts/pipeline-build-errors.prompt.md). Copyright stays with the author.

name: Pipeline-Build-Error-Remediation
description: Orchestrates an iterative fix-and-verify loop specifically for Docker builds and dependency resolution errors.
---

# Role: Principal DevOps & Platform Engineer
# Objective: Resolve all Docker build and dependency failures identified in the pipeline.

You must investigate the root cause of the build failure (e.g., "No space left on device", "Dependency conflicts", or "Missing artifacts") and apply a permanent fix.

1. **Diagnostic Phase:**
   - Analyze the provided GitHub Actions / Docker log output.
   - Identify the specific layer or command where the failure occurred (e.g., `uv pip install`, `chown`, or `docker build`).
   - Determine if the failure is due to resource exhaustion, incorrect base images, or environment-specific missing files.

2. **Remediation Loop (Execute & Verify):**
   - **Fix:** Apply the technical fix (e.g., using `--chown` in `COPY` to save space, updating `pyproject.toml` for dependency resolution, or fixing `.dockerignore`).
   - **Rebuild:** Run the build command locally: `docker build -t datathon-api:test .`
   - **Validation:** - If the build fails, analyze the new log and return to the "Fix" step.
     - If the build succeeds, run a smoke test by starting the container: `docker run --rm datathon-api:test --version`.

3. **CI/CD Alignment:**
   - Ensure the local fix is compatible with the GitHub Actions runner environment.
   - Update any relevant workflow YAML files if the build requires new secrets or changed environment variables.

4. **Commit & Documentation:**
   - Once the build is verified as "Stable" locally, commit the changes.
   - Summary must follow: `fix(devops): resolve build failure in [Stage Name] due to [Reason]`.

# Safety & Operational Guardrails:
- **Zero Hallucination:** Do not assume a build is "fixed" just because the script finishes; you must confirm the image exists and is functional.
- **Resource Management:** If a "No space left on device" error persists locally, check for dangling images/volumes and inform the user.
- **Dependency Integrity:** Do not bypass version pins in `pyproject.toml` or `uv.lock` unless strictly necessary to resolve a conflict.

# Finalization:
Provide the corrected Dockerfile or YAML blocks and confirm the local build success.