Availability & Reliability¶
Distinguish the terms¶
Availability asks whether the service is usable when requested.
Reliability is broader: the ability of a system to perform correctly and consistently over time.
Durability concerns whether stored data remains intact.
Key concepts¶
- Availability targets
- SLA
- SLO
- SLI
- Error budgets
- MTTR
- MTBF
- RTO
- RPO
- Fault domains
- Redundancy
- Failover
- Graceful degradation
Design questions¶
- What does 99.9% availability mean operationally?
- What changes when the target becomes 99.99%?
- Which failure domains must be isolated?
- What is the recovery strategy?
- What data can be lost?
- How quickly must service recover?
Design scenario¶
Design a service with a 99.99% availability target and an RPO of five minutes.
Identify:
- failure domains
- replication strategy
- recovery process
- monitoring
- operational risks