Production infrastructure & observability
Several production websites and APIs on a single, well-monitored server — automatic HTTPS, shared databases, object storage, and full tracing, logs and metrics.
The challenge
Small businesses need production-grade reliability without enterprise cloud bills — and without a dedicated operations team.
The solution
A lean, container-based setup on one VPS that hosts multiple products and domains:
- Reverse proxy — Caddy with automatic HTTPS certificates through Cloudflare, serving static sites and APIs for several domains.
- Shared services — a TimescaleDB / PostgreSQL server with a separate database per product, and self-hosted S3-compatible file storage.
- Repeatable deploys — scripts with dry-run mode, automatic backups before every change, database migrations and health checks.
- Separate environments — isolated development, staging and production.
Observability
- Tracing — OpenTelemetry in the backends and in the browser, stored in Grafana Tempo.
- Logs — structured logs collected by Promtail into Loki.
- Metrics — Prometheus with host, container and database exporters.
- Dashboards & alerts — Grafana, plus Telegram alerts for critical business events.
- Runs anywhere — including a memory-tuned version of the monitoring stack on a Raspberry Pi.
