Managing the resilience/cost balance
In the practice "3B1 - Template Management and Architecture Best Practices", we talk about the resilience/cost balance. But to go further, it is also necessary to review its Disaster Recovery strategy for each application and adapt it to the SLA according to the best compromise.
- The Backup & Restore is the cheapest, perfectly adapted to applications that have a RTO/RPO of a few hours
- The Pilot Light, it's when you have for example a "mirrored" database but with VMs off, a bit more expensive than a Backup & Restore it works for most applications that do not have extreme SLA requirements (less than 1h)
- The Warm Standby, it's when the VMs are already running but with limited scalability, we are almost in real time but potentially in slightly degraded quality if incidentally
- The Hot Standby Multi-Site, there we are totally resilient for real time SLAs. No loss of service is tolerated, but of course this has a cost.