We deployed over the years several fully redundant Docker Swarm HA cluster across multiple Hetzner nodes. Key features:
- Keepalived for floating IP failover across 3 control-plane nodes
- GlusterFS for replicated storage of persistent container data
- Graylog for real-time log aggregation
- Netdata + Zabbix for resource monitoring, alerting, and trend analysis
- Docker stacks for Laravel, WordPress, Redis, MariaDB Galera, Mailu, and more
- Auto-deploy pipelines via GitLab CI/CD
- Unified UFW firewall with automated Ansible hardening
Result: instant failover, 100% data replication, and easy scalability.