"Hello, World" (the longer part)

This time, we'll manually do all the steps, avoiding Maven as possible. This is made so you could easily understand the savings that Lambada introduce to you.

Scaffolding it:

Create a blank maven project:

$ mvn archetype:generate -DgroupId=myorg -DartifactId=hello-lambda \
  -Dversion=0.0.1-SNAPSHOT -Dpackage=myorg

Then, when prompted, simply type return all the way. Import this project. Lets start by writing the function signature:

package myorg;

public class MyLambda {
  public String sayHello(String in) throws Exception {
    System.out.println("in: " + in);

    return "{\"hello\": \"world\"}";
  }
}

Also:

• delete App.java and AppTest.java
• remove the dependencies section```

Then, compile it:

$ mvn package

Now we'll emulate the deploy part of it. Replace <mybucket> with a bucket name you picked.

$ aws s3 mb s3://<mybucket>
$ aws s3 cp target/hello-lambda-0.0.1-SNAPSHOT.jar s3://mybucket/
$ aws iam list-roles | jq -r .Roles[].Arn | grep lambda_basic_execution 
# Pick your ARN for lambda_basic_execution
$ aws lambda create-function --function-name say-hello --runtime java8 --role arn:aws:iam::<youraccountid>:role/lambda_basic_execution --handler myorg.MyHandler::sayHello --code S3Bucket=<mybucket>,S3Key=hello-lambda-0.0.1-SNAPSHOT.jar
{
    "CodeSha256": "8a9SGPDW6Yy0HWwVUWDLaKEg3OJv/oEACp9uP/60sg4=", 
    "FunctionName": "say-hello", 
    "CodeSize": 2067, 
    "MemorySize": 128, 
    "FunctionArn": "arn:aws:lambda:us-east-1:235368163414:function:say-hello", 
    "Version": "$LATEST", 
    "Role": "arn:aws:iam::235368163414:role/lambda_basic_execution", 
    "Timeout": 3, 
    "LastModified": "2016-04-12T12:54:50.008+0000", 
    "Handler": "myorg.MyHandler::sayHello", 
    "Runtime": "java8", 
    "Description": ""
}

There, your function is created! However, a new deploy needs to use a different call: update-function-code:

$ aws lambda update-function-code --function-name say-hello --runtime java8 --role arn:aws:iam::235368163414:role/lambda_basic_execution --handler myorg.MyHandler::sayHello --code S3Bucket=mylambda-bucket-aldrin,S3Key=hello-lambda-0.0.1-SNAPSHOT.jar

Of course, you can also change the configuration of the function itself from the console.

Plus, if you use dependencies, they aren't included, requiring you to build a fat jar instead.

Enter Lambada

Lambada + beanstalker automates this process by:

• @Lambada annotation defines the function code
• lambada:generate collecting those annotations and generating a lambda-definitions.json file
• beanstalker:upload-source-bundle creates (if needed) then uses an S3 Bucket which is always reserved and available for your user account. All it takes is to declare the property beanstalk.s3bucket to empty.
• The lambda:deploy-functions reads the output of lambada:generate and replaces your role ids where applicable then figures out which functions to call (create-function, update-function and update-function-code)

Understanding the Maven Plugins

From now on, we'll refer the Maven Plugins to their real name: Mojos.