What is concurrency in AWS Lambda?
Concurrency in AWS Lambda refers to the number of instances of a function that can run simultaneously. When multiple events invoke a Lambda function at the same time, AWS Lambda launches multiple instances (up to a limit) to handle the load in parallel. This helps in handling scale and latency effectively for event-driven workloads.
- 1
Unreserved Concurrency: The default concurrency pool shared by all functions in an account.
- 2
Reserved Concurrency: Dedicates a fixed number of concurrent executions to a specific function.
- 3
Provisioned Concurrency: Pre-initializes Lambda instances so they’re ready to respond instantly, ideal for low-latency needs.
- 1
Each function can scale up to the regional concurrency limit by default (e.g., 1,000 concurrent executions).
- 2
Setting reserved concurrency for a function prevents other functions from affecting its concurrency.
- 3
Provisioned concurrency helps eliminate cold starts for high-performance applications.
Scenario Questions
0-2 years experience
- 1You need to write a Lambda that processes S3 upload events and you notice occasional delays. How would you use concurrency settings to improve its responsiveness?
- 2If a function has no reserved concurrency configured and you see throttling errors, what’s happening and what quick fix would you apply?
2-5 years experience
- 1During a rollout you observe that a spike in traffic is causing downstream API timeouts. Walk me through how you’d diagnose whether Lambda concurrency is the root cause and what you’d change.
- 2Your team wants to guarantee low latency for a high‑traffic endpoint. How would you decide between increasing reserved concurrency versus enabling provisioned concurrency?
5-8 years experience
- 1Design a strategy for a multi‑tenant SaaS platform where each tenant’s Lambda must have isolated concurrency limits while still sharing a global account limit.
- 2You’re seeing bursty traffic that overwhelms a downstream database. Explain how you’d use concurrency controls, throttling, and back‑pressure to protect the database without sacrificing Lambda throughput.
8+ years experience
- 1At an enterprise scale you need to migrate dozens of existing Lambdas to use provisioned concurrency without causing service disruption. Outline the migration plan, including testing, rollout, and rollback steps.
- 2How would you architect a cross‑team governance model for Lambda concurrency quotas that balances cost, performance, and fairness across many product lines?
Follow-up Questions
- How would you set up alerts for throttling events?
- What trade‑offs exist between reserved and provisioned concurrency?
- Can you describe a scenario where you’d need to adjust concurrency limits on the fly?