We use analytics and advertising cookies to understand how the site is used and whether our ads on Facebook and Instagram work. They are set only if you accept. See our Privacy Policy for details.
Lambda questions in cloud, backend and DevOps interviews are rarely "what is serverless". They are scenarios: a function throttles under an SQS burst, a VPC-attached function times out calling an API, a payment handler runs twice. Interviewers want to hear how concurrency, event source mappings and retries actually behave.
Each question below is a real scenario with four plausible answers and an explanation of why the wrong ones are wrong.
Answer a real AWS interview question before you browse the Lambda list. No signup needed.
A content management application runs on EC2 instances in an Auto Scaling group across three Availability Zones. Every instance must read and write the same set of user-uploaded files through a standard file system interface (POSIX), and files must remain available if an AZ fails. Which storage solution should a solutions architect choose?
Practice in a session
10 questions from AWS, one after another, then your score and streak.
Concurrency (reserved vs provisioned, account limits and throttling), cold starts and how to reduce them, event source mappings for SQS, Kinesis and DynamoDB Streams (batching, partial batch failures), retry behaviour and on-failure destinations or DLQs for asynchronous invocations, the 15-minute timeout, the memory-to-CPU coupling, VPC attachment and NAT, execution roles, and idempotent handlers.
Reserved concurrency sets aside part of the account's concurrency for one function, which both guarantees it that much and caps it there; it costs nothing and does not pre-initialise anything. Provisioned concurrency keeps a number of execution environments initialised so requests skip the cold start, and you pay for it while it is configured.
Keep the deployment package and init code small (lazy-load what the first request does not need), give the function more memory (which also gives it more CPU), use provisioned concurrency for latency-critical paths, and use SnapStart on supported runtimes (Java, Python, .NET) to restore from a snapshot of the initialised environment.
For asynchronous invocations Lambda retries twice by default, then sends the event to an on-failure destination or a dead-letter queue if one is configured. With an SQS event source a failed batch becomes visible on the queue again after the visibility timeout; reporting batch item failures retries only the failed messages, and the queue's redrive policy moves a message to its DLQ after maxReceiveCount attempts.