Availability zones are separate facilities inside one region, close enough for synchronous replication, so spreading across them protects you from a rack, power or facility failure with little added complexity. Multi region protects you from a whole region being degraded and can satisfy data residency, but replication becomes asynchronous, so you accept data loss on failover and take on real cost and operational complexity. Most workloads need multi zone plus cross region backups.
Why interviewers ask this
This question checks whether you match resilience investment to actual requirements instead of reaching for the most impressive architecture. The interviewer wants honesty about the cost of multi region: asynchronous replication, data transfer, split brain risk, and the fact that an untested failover path does not work. Recommending backups in a second region as the pragmatic middle ground is usually the answer they want.
How to structure your answer
- Define both and the failure each one addresses.
- Explain why replication mode differs and what that costs you.
- State the pragmatic default for most workloads.
- Say what would make you actually go multi region.
Example answer
Zones are separate physical facilities within a region with fast private links between them, so you can replicate synchronously and lose a zone without losing data. Spreading across three zones is close to free in complexity terms and it covers the failure mode you are most likely to hit. Regions are geographically distinct, so going multi region protects you when an entire region is degraded, and it is sometimes required for data residency. The honest cost is that replication across regions is asynchronous, so failing over means accepting some data loss, and you now have to solve routing, split brain, cross region data transfer charges and drift between two stacks that must stay compatible. For most workloads I recommend multi zone for availability plus backups and machine images replicated to a second region for recovery, which gives a recovery time of hours rather than minutes at a fraction of the cost. I would only build active multi region when the business has put a number on an hour of downtime that justifies it, and then I would insist on regular tested failovers, because an untested failover is a hope.
Walking into this interview soon? GhostPilot listens to your live call, spots the question the moment it is asked, and puts a structured answer on your screen in real time. Try it on your next mock, or grab a $29 Session Pass, no subscription, for the real thing.
See how it worksFollow-up questions to expect
- How would you handle a database that cannot replicate synchronously across regions?
- How do you keep two regional stacks from drifting apart?
- What is your plan if the region outage affects the control plane you need to fail over?
Related cloud engineer questions
Your interviewer will ask their own version of this. Paste your actual job description into the free Question Predictor and get the 20 questions that role is most likely to ask, with what each one is really probing.
Predict my questions