The Best of Series
momjian.us/presentations
The presentation above addresses requirements open source (mostly free solutions) for PostgreSQL functionality to:
– High Availability – No application reconfigure after fail-over/switch – RepMgr (an open source tool for managing PostgreSQL replication and failover)
– RPO (recovery point objective / maximum amount of committed data lost) = 0 after fail-over/switch-over
– Fail-over should be automatic – switch over is ALWAYS manual / planned – Uses HAProxy to Route with Keepalived daemon to manage virtual IP that always routes to the active node
– Back-up and recovery – Barman (Python based tool) to backup and recover clusters and databases
– Automatic deployment and configuration – Terraform / Ansible
– Monitoring – Prometheus / Grafana
• Prometheus (https: / /prometheus.io)
• Grafana (https: //grafana.com)
• Scalable
• Collectors for Postgres, OS, Keepalived, HAProxy
(https: //github.com/wrouesnel/postgres _exporter)
(https: / /github.com/prometheus/node _exporter)
(https:/ /github.com/cafebazaar/keepalived-exporter)
(no separate exporter is needed for HAProxy Metrics)
– Encryption TLS in transit / at rest
– Auditing
– Automatic maintenance
– Performance insights in queries