Adapting Java for the Serverless world
summary
Shows concrete numbers and recipes: Java cold starts often reach seconds not milliseconds, simple Java + DynamoDB hit 10–20s, optimizations (AWS SDK v2, async HTTP, static init tuning, tiered-compilation off) cut that dramatically. Demonstrates building GraalVM native image deployed as AWS custom runtime, compares startup and memory for Quarkus/Micronaut/Spring Boot, and uses AWS Sam and Docker for builds. Provides measured results (cold starts down to ~0.5–0.8s for native images), trade‑offs (long build times, CI cost) and when to prefer provisioned concurrency or containers versus AOT.
AI-written summary · about
chapters (7)
- 0:00Start
- 8:08Creating AWS Lambda with Java 1/3
- 21:13Best Practices and Recommendations
- 29:03Lambda Power Tuning 2/2
- 43:12AWS Lambda Deployment of Custom Runtime with SAM
- 49:45Spring Native Example
- 53:34Lambda demo with common Java application frameworks
related talks
51mAdapting Java for the Serverless world!
56mAfraid of Java cold starts in Serverless? Fear not, Java is fast
49mServerless Java Web Applications on AWS Lambda with Micronaut Quarkus & Spring Boot
53mHigh performance Serverless Java on AWS
17mAdopting Java for the Serverless world
46mReduce Cold Starts for Java Serverless Applications in AWS
56mThinking in Serverless Java (Code)
53mAdapting Java for the Serverless world