Free sample question

When would you recommend a hybrid or multi-cloud strategy, and when is it a mistake?

Mid-Senior · Architecture · Cloud Strategy · Chapter 1: Multi-Cloud & Hybrid Cloud Strategy, question 1 of 13 · from Cloud Interview Mastery: AWS, Azure & GCP

What the interviewer is really testing

whether your architecture decisions start from business drivers like residency, negotiating position, or best-of-breed services, and whether you have the judgment to recommend a single cloud when no driver exists.

The 30-second answer

Hybrid means pairing on-prem infrastructure with public cloud; multi-cloud means running more than one public provider. I'd only recommend either with a concrete driver: a regulator demanding data residency, a workload pinned to legacy hardware, or one provider's clearly better service: BigQuery for analytics is the classic case. The default for most companies is one primary cloud, because every extra platform multiplies IAM models, networking, and on-call skills. And cross-cloud traffic rides internet egress at roughly $0.09/GB on AWS, so I'd colocate compute with data from day one.

Follow-ups the interviewer will probe

How would you actually run Kubernetes across AWS and GCP with failover?
Separate EKS and GKE clusters built from shared Terraform modules, stateless services deployed to both through GitOps (Argo CD), and a provider-neutral global load balancer with health checks steering traffic. The hard part is state: active-passive Postgres with cross-cloud replication, or CockroachDB if both sides must write. I'd test failover on a schedule, not after the outage.
Doesn't Kubernetes make the clouds interchangeable?
For stateless compute, mostly. The container and manifest move; everything they touch does not: load balancer annotations, storage classes, and IAM bindings (IRSA versus Workload Identity) are all provider-specific. Data is the real anchor: a multi-terabyte warehouse doesn't relocate over a weekend. I treat Kubernetes as reducing switching cost, not eliminating it.
How do you keep multi-cloud spend visible and under control?
Colocate compute with data so cross-cloud chatter stays thin, then unify reporting: all three major providers now publish billing data in the FOCUS open format, so one FinOps pipeline can compare them. I'd alert on egress line items specifically; they grow silently, and committed-use discounts don't cover them.

Recall hook

“Multi-cloud is a driver, not a default.”

go hybrid for residency and legacy, add a second cloud only for a service that clearly wins, and stay on one cloud when neither applies.

What the book adds to this question

In the ebook every question runs three pages. Between the 30-second answer and the follow-ups it adds a deep dive with a diagram, a decision framework, and a pitfalls-and-signals table. Cloud Interview Mastery: AWS, Azure & GCP has 50 questions across 8 chapters and includes the free Interview-Day Playbook.

Buy this ebookSee all 50 questions and prices

Want full three-page questions? Download the free 8-question PDF sample (from Cloud Interview Mastery).

Sample questions from the other books