Using a layer for lambda with Quarkus JVM (Non-native)

Question

Maybe someone help me with a question.

I'm trying to use Quarkus JVM to deploy lambda functions in AWS. But I cannot set a layer using the function.zip genereated by Quarkus when I use "mvn clean package".

When I set directly on each function, "function.zip" works.

Like the following:

Globals:
  Function:
    Tracing: Active
    CodeUri: target/function.zip
    Handler: io.quarkus.amazon.lambda.runtime.QuarkusStreamHandler::handleRequest
    Runtime: java17
    Timeout: 30
    MemorySize: 1024

But I would like to use a layer resource and use it the layer in the lambda function.

For example:

LambdaLayer:
  Type: AWS::Serverless::LayerVersion
  Properties:
    LayerName: Quarkus-Layer
    Description: QuarkusLayer
    ContentUri: target/function.zip
    CompatibleRuntimes:
      - java17
    CompatibleArchitectures:
      - x86_64
    LicenseInfo: MIT

And I reference that layer in the lambda resource:

Globals:
  Function:
    Tracing: Active
    CodeUri: target/empty-function.zip
    Handler: io.quarkus.amazon.lambda.runtime.QuarkusStreamHandler::handleRequest
    Runtime: java17
    Timeout: 30
    MemorySize: 1024
    Layers:
     - !Ref LambdaLayer

But If I use layers it doesn't work, Can't find quarkus handling class.

And again, when I set direct "function.zip" directly on each lambda or globally it works great. But I feel like it's not efficient to use it that way, I'd like to use layers.

I've searched GitHub for similar projects and they all use directly in each function they don't use layers.

CloudWatchLogs Image