Skip to main content

High Availability & Disaster Recovery (Availability & DR)

Status

Owner: Tech lead · Status: draft · Canonical source: flezi.infra.iac/docs/architecture-snapshot-14-08-2026.md

Details platform redundancy, automated backup schedules, and disaster recovery recovery objectives.

High Availability Architecture​

  • PostgreSQL Flexible Server: Deployed in Zone-Redundant High Availability mode with synchronous standby replication and automated failover (under 120s).
  • Azure Managed Redis: Clustered multi-node setup with automated primary/replica failover.
  • Stateless Compute: Backend and AI Runtime instances run stateless, enabling rapid restart and horizontal scaling.

Backup Schedules & Recovery Objectives​

  • Database Backups: Automated daily full backups with point-in-time restore (PITR) up to 35 days.
  • RPO (Recovery Point Objective): < 5 minutes for database transactions.
  • RTO (Recovery Time Objective): < 60 minutes for complete infrastructure reconstruction from Terraform.

Existing Controls​

  • Zone-redundant database operational on production tiers.
  • Infrastructure fully reproducible from IaC definitions.

Evidence (Files & Configuration)​

  • flezi.infra.iac/envs/prod/database.tf
  • flezi.infra.iac/ansible/

Limitations & Roadmap​

  • Scheduled annual disaster recovery simulation drills.