Building the OpsByFabian production platform
Private deployment, monitoring, backups and tested recovery.
Internal infrastructure
The problem
Software isn’t finished when it deploys. I wanted a production environment I could administer privately, monitor from outside and restore when something failed.
The approach
Build an Ubuntu VPS environment with hardened access and private Tailscale administration. Add container deployment, HTTPS, databases, automation and monitoring. Then test reboots, exposure, backup integrity and actual application restores.
What the work covers
Private administration
Hardened SSH, Tailscale and IPv4/IPv6 firewalling. External port checks and firewall persistence tests validate the intended access boundaries.
Deployment and services
Docker, Coolify and Traefik for application deployment, DNS and TLS. PostgreSQL and Redis support the application and automation workloads.
Operational visibility
External monitoring, infrastructure monitoring, error tracking, analytics and change detection. Service health checks and backup heartbeat monitoring belong to the operations layer.
Recoverable backups
Encrypted Restic backups to Cloudflare R2, plus provider-level backups, database and volume backups. Recovery tests cover the data and the application together.
Automation and AI environment
n8n, Evolution API and a remote Hermes AI agent environment, with their data and recovery requirements included in the platform.
Tested beyond installation
Exposure and restart
External port exposure checks, firewall persistence, a complete VPS reboot and automatic service recovery.
Detection and integrity
External outage detection and encrypted backup integrity checks.
Hermes restore
A full restore of the Hermes environment.
n8n recovery
PostgreSQL restore into an isolated database, restoration of the data volume, a temporary restored application and successful authentication into that instance.
Internal OpsByFabian platform. This describes the build and validation supplied by Fabian for September 2026. Availability, support coverage and recovery objectives are scoped per client; high availability and multi-region operation are not claimed.
What isn’t working in your business?
Tell me about the workflow, the tools or the idea. We’ll work out a sensible next step.