AWS Lambda Handler vs Controller?

[ghost]

So I came across this issue (#22595) while troubleshooting another problem, and became confused by the reference to the "controller" method vs "handler" method.

From my understanding, everything in AWS Lambda has to use a handler approach. Quarkus wraps your handler in their QuarkusStreamHandler wrapper class to initialize quarkus, which then passes off the the AmazonLamabdaRecorder to scan your classpath and look for your own implementation of RequestHandler. If you're using the quarkus-amazon-lambda-http or quarkus-amazon-lambda-rest extensions, they include a handler implementation for you. From my own testing, if you try to make a recursive call in your local env, these handlers will block until the initial request errors out, confirming their ability to handle a single request per instance.

Am I missing something here? I'm not sure what the reference to "controller method" is in that issue, but I've reviewed 100% of the very limited documentation on quarkus lambda extensions and haven't found anything helpful. There also seems to be some concerns in that thread by one of the quarkus contributors about the native compiled app would only be able to support a single request at a time:

But based on my limited understanding of lambda, that's simply how it works. A single instance with a single handler can only ever handle a single request at a time. From AWS Lambda documentation:

Perhaps I'm just an idiot and I'm missing something very simple, but this has really thrown me for a loop. Anyone have any insight?

[quarkus-bot[bot]]

/cc @matejvasek (amazon-lambda), @patriot1burke (amazon-lambda)