Harness Continuous Verification Example Application

Thanks for checking out the Harness Continuous Verification Sample Application. This application is designed to be run in conjunction with a Prometheus Health Source with Harness.

What The Example Application Does

• Generates Metrics to default Micrometer [Spring Boot] Prometheus Registry.
• Has a 'stable' and 'unstable' version. Can be triggered via a Docker Tag e.g during a deployment or on-the-fly via the REST endpoints.

Modifying Application

Can modify this application to test out your needs. A few items that are needed.

• OpenJDK 19 - brew install openjdk
• Maven 3 - brew install maven

Running Locally

mvn spring-boot:run

Can toggle explore the endpoints:

• localhost:8082
• localhost:8082/actuator
• localhost:8082/actuator/prometheus
• localhost:8082/run-unstable
• localhost:8082/run-stable

Docker Build

Can have two tags, for stable and unstable. Can change the entry point [comment/un-comment] which first metric to use -> main/src/main/java/io/harness/cv/example/app/CvExampleApp.java and re-build and re-push to a Docker Registry of your choice.

• Docker Install - brew install docker

docker build --platform linux/amd64 -t rlachhman/cv-example-app:stable .
docker push rlachhman/cv-example-app:stable  

docker build --platform linux/amd64 -t rlachhman/cv-example-app:unstable .
docker push rlachhman/cv-example-app:unstable  

Deploying to Kubernetes with Harness

Can leverage the harness-cv-example-deployment.yaml and values.yaml to deploy in Harness to pick the specific stable or unstable tag. The Deployment YAML also creates a LoadBalancer Service. With a public cloud this can be a public IP or using something locally like Minikube can run minikube tunnel and head to the public IP e.g kubectl get services. This is most likely local host or 127.0.0.1:8082] to view the application.

Deploying to ECS with Harness

Can leverage the ecs_task.json and service_definition.json modified to your AWS resources to deploy to ECS.

Installing Prometheus

Installing Prometheus from Helm is easy.

helm repo add prometheus-community https://prometheus-community.github.io/helm-charts

helm repo update

helm upgrade --install prometheus prometheus-community/prometheus \
--namespace prometheus --create-namespace

For an example, you can also expose over NodePort.

kubectl expose deployment -n prometheus prometheus-server --type=NodePort --name=prometheus-service

Then browse to node_public_ip:nodeport.

e.g

#Node External IP
kubectl get nodes -o wide

#NodePort for prometheus-service
kubectl get svc -n prometheus

Metrics

There are two metrics that can be modified.

• CV_Counter_Example_total
• CV_Gauge_Example{}

In the stable version these are flat. In the un-stable versions these metrics increment up every 15 seconds. Can modify src/main/java/io/harness/cv/example/app/GenerateUnstableMetrics.java and change the amount, time, or add another Actuator Metric which will be wired to Prometheus for you.