How to develop, run and optimize Spring Boot 3 application on AWS Lambda

on YouTube ↗44m

" summary

Compares four deployment paths: AWS Serverless Java Container, AWS Lambda Web Adapter, Spring Cloud Function and custom Docker, measuring cold starts, warm latencies and tradeoffs. Baseline cold P90 sits around 4–7s; warm P90 ≈10ms. Enabling SnapStart with priming (CRaC pre‑snapshot hooks that hit DynamoDB / warm class paths) drops many cold P90s to ~1s; extra caching improves further. Building a GraalVM native image yields P90 ≈900ms; ARM64 cuts cost ~25% but shows ≈10% slower cold starts.

AI-written summary · about

" chapters (11)

1. 0:00Spring Boot on AWS Lambda
2. 5:41AWS Serverless Java Container
3. 9:12Lambda Web Adapter tool
4. 11:42Spring Cloud Function on AWS
5. 13:04How Java Lambda works
6. 18:21Optimizing cold starts with SnapStart
7. 21:33Priming techniques for SnapStart
8. 28:57Measuring SnapStart gains
9. 33:37GraalVM Native Image for Spring Boot
10. 41:48SnapStart vs GraalVM
11. 43:12Q&A

" related talks

• 53mHigh performance Serverless Java on AWSVadym Kazulkin · GeeCON 2024
• 46mReduce Cold Starts for Java Serverless Applications in AWSVadym Kazulkin · JCON WORLD ONLINE 2023
• 51mHow To Reduce Cold Starts for Java Serverless Applications in AWSVadym Kazulkin · GOTO Amsterdam 2024
• 34mThawing Java on AWS LambdaDaniel Raniz Raneland · Craft Conference 2022
• 49mServerless Java Web Applications on AWS Lambda with Micronaut Quarkus & Spring BootVadym Kazulkin · JCON EUROPE 2026
• 19mTen things that improve the performance of your Java based AWS LambdaJeroen Reijn · J-Fall 2022
• 46mThawing Java on AWS Lambda Reducing Cold Start Times from 11 Seconds to 1Daniel Raniz Raneland · JCON WORLD ONLINE 2023
• 01h 00mAdapting Java for the Serverless worldVadym Kazulkin · Voxxed Days Luxembourg 2022