Deployment

What you need

• A PostgreSQL database (any managed provider works).
• A deployment platform that runs a Node.js server.
• A small set of required environment variables, validated at startup — a missing one fails fast with a clear message rather than a confusing downstream error.

Health checks

A dedicated health endpoint checks database connectivity and returns a simple ok/unavailable status, without ever revealing hostnames, connection strings, or version numbers. Point your platform’s health check at it.

What’s honestly not automated yet

There is no background job scheduler in this codebase today. Concretely, that means:

• Configured data retention windows are not automatically enforced by a scheduled deletion job.
• Outbound webhook delivery is immediate and best-effort with a couple of bounded retries, not a durable queue — a delivery lost during a restart is logged as a gap, not silently dropped, but it isn’t retried later.

If your deployment needs either of those to run on a real schedule, that requires adding scheduler infrastructure on top — we don’t pretend one exists.

Backups

Database backup strategy is entirely deployment-dependent. Use your Postgres provider’s backup or point-in-time-recovery feature — don’t assume backups exist just because a managed database is in use.