Interview kitsBlog

Your dream job? Lets Git IT.
Interactive technical interview preparation platform designed for modern developers.

XGitHub

Platform

  • Categories

Resources

  • Blog
  • About the app
  • FAQ
  • Feedback

Legal

  • Privacy Policy
  • Terms of Service

© 2026 LetsGit.IT. All rights reserved.

LetsGit.IT/Categories/Cloud
Cloudhard

How do you design for high availability across failures (multi-AZ vs multi-region)?

Tags
#high-availability#multi-az#multi-region#failover
Back to categoryPractice quiz

Answer

Multi-AZ protects you from a datacenter outage inside a region with lower latency and simpler ops. Multi-region can survive a full region outage but adds latency, data replication complexity, and higher costs.

Advanced answer

Deep dive

High availability starts with defining failure domains and objectives:

  • **RTO** (how quickly you must recover)
  • **RPO** (how much data loss is acceptable)

Multi-AZ (within a region)

Typical baseline:

  • stateless app instances spread across 2–3 AZs,
  • load balancer with health checks,
  • managed DB in multi-AZ mode or replication/failover,
  • redundant caches/queues.

Multi-region (across regions)

Used for disaster recovery or global latency:

  • **active-passive**: one region serves traffic; the other is standby (simpler, but failover time).
  • **active-active**: both serve traffic (harder: data consistency, conflict resolution, split-brain risks).

Practical guidance

  • Do multi-AZ by default for production.
  • Add multi-region when your RTO/RPO or regulatory needs require surviving a region outage.

Common pitfalls

Related questions

Cloud
Disaster recovery: backup/restore vs warm standby vs active-active — what’s the trade-off?
#disaster-recovery#failover#multi-region
Microservices
Multi-region microservices: what are the main benefits and the main pain points?
#microservices#multi-region#availability
  • Ignoring data replication semantics (eventual consistency, write conflicts).
  • No automated failover testing.
  • Treating “multi-region” as only DNS changes; the data layer is usually the hard part.