From Single Host to Distributed Overhead. Of course it runs FreeBSD!
Building a multi-continent FreeBSD infrastructure with modern GitOps practices
From a single server to a multi-continent overhead. Of course it runs FreeBSD!
Series Overview
How I started with a single VM and ended with multi-locality replicated overhead.
Planned Posts
Post 1: "From a Single Host to Global Distribution"
- The evolution from tesla-only to "6"-host distributed setup
- Geographic distribution rationale
- High-level service overview
Post 2: "Wireguard Mesh - Connecting (not only) Continents Securely"
- Multi-region network design
- Host-to-host and jail-to-jail connectivity
- Security considerations and key management
- Performance across geographic distances
Post 3: "Database High Availability with PostgreSQL and Patroni"
- Cross-region PostgreSQL clustering
- Patroni configuration for geographic distribution
- HAProxy integration and endpoint management
- Backup and disaster recovery strategies
Post 4: "Distributed Storage with Garage S3"
- Why Garage over alternatives
- Multi-region object storage setup
- Integration with Nextcloud and other services
- Performance and consistency considerations
Post 5: "Mail Infrastructure - Building Reliable Email Services"
- Postfix/Dovecot/Pigeonhole/rspamd stack
- Multi-region mail handling strategies
- Dovecot replication mechanics
- rspamd integration with Redis/Valkey
Post 6: "Monitoring Everything - Prometheus, Loki, and Beyond"
- Centralized monitoring with Prometheus/Grafana
- Log aggregation with Loki and Promtail
- Uptime-kuma for endpoint monitoring
- Discord integration for alerting
Post 7: "GitOps and Automation with Gitea Act Runners"
- Custom orchestration development
- Distributed automation across hosts
- Package building and distribution
- Future orchestration goals