railway-devops

# Railway DevOps Skill This skill enforces the strict operational rules for CBB Edge production environments on Railway, preventing unresolvable hostname errors and ensuring reliable log filtering. ## When to Use Activate this skill when: - Running Python scripts that connect to the production database. - Tailing or filtering Railway logs. - Verifying environment variables. - Investigating `OperationalError: could not translate host name` failures. ## Core Workflow ### 1. Database Command Rule (CRITICAL) Production `DATABASE_URL` uses internal Railway hostnames that do NOT resolve locally on Windows. | Scenario | Command Pattern | Why | |----------|-----------------|-----| | **Local / No DB** | `railway run <cmd>` | Injects env vars only. Safe for logs/local logic. | | **Production / DB Access** | `railway ssh <cmd>` | Runs inside the container where DNS resolves. | ### 2. Log Tailing & Filtering Use the following patterns to monitor the production service: ```bash # Tail all logs railway logs --follow # Filter logs for a specific job (e.g., player_id_mapping) railway run python scripts/devops/railway_logs_filter.py --job player_id_mapping --lines 50 ``` ### 3. Environment Variable Audit ```bash # List all variables to check for missing keys railway variables # Set a critical gate variable railway variables set INTEGRITY_SWEEP_ENABLED=false ``` ## Guidelines - **No Local DB Connections:** Never attempt to connect to the internal production DB hostname directly from a local terminal. Always use `railway ssh`. - **Fail-Safe Triage:** If a command fails with a "host name" error, retry with `railway ssh`. - **Selective Logging:** Prefer `railway_logs_filter.py` over raw `railway logs` for specific job debugging. ## Examples "Run the database health check script." -> *Action:* `railway ssh python scripts/devops/db_health.py` "Check the logs for the ingestion orchestrator." -> *Action:* `railway run python scripts/devops/railway_logs_filter.py --job ingestion_orchestrator`