Production infrastructure

It proves the production claim.

I am not deploying to a platform that owns the reverse proxy for me. The nginx configuration with its rate limits and logging, the per-site virtual hosts, the Redis configuration and three systemd units — each with an installer script — are files I wrote and can show.

Three services, three units.

The site, the API and the Celery worker each get their own systemd unit and their own installer, so they start, stop and fail independently. Readiness is reported through sdnotify, so systemd knows when a service is actually serving rather than merely running — which is the difference between a restart that waits and a restart that drops requests.

Highlights

• nginx, Redis and three systemd units under version control, with installers
• sdnotify readiness rather than a guess at startup time
• Rate limiting and per-site logging configured at the proxy
• Fifteen files that describe the production machine completely

Stack

nginxsystemdRedisRaspberry Pi