Cloud Platforms Interview Questions & Answers

An IAM Role is like a temporary badge: it defines a set of permissions that someone can borrow for a while, rather than being tied permanently to one identity. That's exactly what an outside user (like someone in a mobile app) needs, so AWS STS issues the short-lived credentials, but it's the Role itself that's being granted and assumed.

Manually SSHing into 100 machines one at a time doesn't scale, and rebuilding every VM is overkill for a small update. The Custom Script Extension lets Azure push and run the same script across many VMs at once through its own control plane, so one action handles the whole fleet instead of a hundred separate logins.

A Cartesian product happens when a JOIN is missing a proper condition, so every row in one table gets matched with every row in the other, producing a huge, mostly-useless result set that costs a fortune to process. Checking totalBytesProcessed shows you when a query scanned way more data than it should have, and fixing the JOIN's ON clause stops the row explosion at the source.

'11 nines' of durability means the odds of ever losing your data are astronomically small, and Amazon S3 gets there by automatically storing multiple copies of every object across different data centers in a region. EBS, EFS, and RDS are built to be fast and available, but they aren't engineered for that same extreme level of data durability.

This specific error means the .NET app crashed while trying to start up inside the web server process, but the error message itself doesn't say why. The real reason (an unhandled exception, a missing config, etc.) shows up in the app's own stdout/stderr output, which App Service captures in its filesystem-based Application Logging, so that's the first place to look before anything else.

Active-active means both regions can accept writes at the same time, which creates a real problem: what happens when the same row gets edited in two places at once? Multi-primary databases like YugabyteDB and CockroachDB are specifically built to detect and resolve those conflicts, something a simple read replica or periodic backup was never designed to handle.

A static website is just files, no server-side code needed, so you don't need a server at all. Azure Blob Storage can serve those files directly over HTTP, which is why it's the cheapest option here, whereas App Service, a VM, or Functions all spin up compute you'd be paying for and don't need.

A server only has a limited number of outbound connection 'slots' (ports) available at once. If traffic spikes and connections open faster than they close, the server can run out of those slots entirely, so new requests just time out, even though the CPU itself has nothing to do and looks perfectly healthy.

Terraform state is just a file that tracks what resources it's already created, and it needs one consistent home so it doesn't get out of sync. Terraform Cloud gives you a single backend that works for both AWS and Azure, with locking and version history built in, instead of juggling two separate, differently-secured storage locations.

High Availability means having enough redundant copies of your app spread across different data centers (zones or regions) that if one goes down, traffic just flows to the ones that are still up. Elasticity and scalability are about handling more LOAD, not about surviving a failure, which is a different problem.