AWS Lambda Java GraalVM with AWS CRT client

Sample AWS Lambda application with GraalVM and AWS CRT client writing to Amazon DynamoDB.

Load test results

0.. warm start

1.. cold start (init duration + duration)

Prerequisites

• AWS Account
• Docker
• AWS SAM CLI
• Artillery for load testing

Setup

1. Create the bundling image:

docker build -t al2023-graalvm:maven .

2. Build the application via the bundling image:

sam build --build-image al2023-graalvm:maven --use-container

3. Deploy the application

sam deploy ## --guided for the first run

Test the application

Get the API-Gateway Url:

 export API_GW_URL=$(aws cloudformation describe-stacks --stack-name <YOUR-STACK_NAME> | jq -r '.Stacks[0].Outputs[] | select(.OutputKey == "RequestHeaderGraalApi").OutputValue')

Do a sample request:

curl -XPOST "$API_GW_URL/request-headers-graal" --header 'Content-Type: application/json'

Run a load test

Run the test via:

artillery run -t $API_GW_URL  -v '{ "url": "/request-headers-graal" }' loadtest.yml

Retrieve results in Cloudwatch-Log insights via:

 filter @type = "REPORT"
    | parse @log /\d+:\/aws\/lambda\/(?<function>.*)/
    | stats
    count(*) as invocations,
    pct(@duration+coalesce(@initDuration,0), 0) as p0,
    pct(@duration+coalesce(@initDuration,0), 25) as p25,
    pct(@duration+coalesce(@initDuration,0), 50) as p50,
    pct(@duration+coalesce(@initDuration,0), 75) as p75,
    pct(@duration+coalesce(@initDuration,0), 90) as p90,
    pct(@duration+coalesce(@initDuration,0), 95) as p95,
    pct(@duration+coalesce(@initDuration,0), 99) as p99,
    pct(@duration+coalesce(@initDuration,0), 100) as p100
    group by function, ispresent(@initDuration) as coldstart
    | sort by coldstart, function