The following is the summary of performance test results collected for the h1-h1-passthrough scenario.

Number of concurrent users: 100
Loop count: 10000
Protocol: https
Ramp-up period: 60s

1. Download and install docker, minikube and Kubernetes.
   • Docker
   • Minikube
   • K8s
2. Start a K8s cluster locally with Minikube

   $ minikube start
   

(Use the docker image of a netty http echo backend from https://hub.docker.com/r/ldclakmal/netty-echo-backend)

1. Run the docker image

   $ docker run -d -p 8688:8688 -e "SSL=true" ldclakmal/netty-echo-backend:latest
   

2. Deploy netty backend to K8s (use the YAML file provided in ballerina-performance/deployment-netty-backend).

   $ kubectl apply -f ./deployment-netty-backend/netty-backend.yaml
   

1. To build test scenario source

   $ bal build
   

   (eg. to build the h1-h1-passthrough program with beta3, navigate to the ./ballerina-slbeta3/h1-h1-passthrough/src and execute $ bal build command)
2. To build and push the docker image

   $ docker build -t <DOCKER_USERNAME>/passthrough_beta3:latest target/docker/passthrough_beta3
   $ docker push <DOCKER_USERNAME>/passthrough_beta3:latest
   

3. To deploy the https passthrough service to K8s

   $ kubectl apply -f ./target/kubernetes/passthrough_beta3  
   $ kubectl expose deployment passthrough-bet-deployment --type=NodePort --name=passthrough-bet-svc-local  
   

4. To test the backend echo service

   $ curl -kv https://localhost:8688/service/EchoService -d '{"size":"50B","payload":"0123456789ABCDEFGHIJKLM"}'
   

5. To test the passthrough service

   $ curl -kv -X POST https://localhost:<NodePort>/passthrough -d '{"size":"50B","payload":"0123456789ABCDEFGHIJKLM"}'
   

6. To run the load test using JMeter

   $ jmeter -n -t ./resources/jmeter/http-post-request.jmx -l results.jtl -Jusers=100 -Jduration=1800 -JrampUpPeriod=60 -Jhost=localhost -Jport=<NodePort> -Jprotocol=https -Jpath=passthrough -Jpayload="./resources/payload/1024B.json" -Jresponse_size=1024