Java has never been a perfect fit for Function as a Service platforms such as AWS Lambda or Azure Functions. While both platforms have official support for Java, Java functions unfortunately suffer from significantly longer cold start times than many other runtimes.

In this talk I will show a simple Spring Cloud Java function running on AWS Lambda with fairly horrible cold start times of around 11 seconds and then show how we can fix this by replacing Spring Cloud with Micronaut and use GraalVM Native Image to perform most of the work during compile time. The end result is cold start times of less than a second, making Java a viable, though not without drawbacks, choice for FaaS platforms that support custom runtimes.

In this updated version I will compare Micronaut and GraalVM with AWS Lambda SnapStart and discuss the different solutions and their specific advantages and drawbacks. If time allows I will also have a look at the new native support in Spring Boot 3.