An overview of health check patterns

Various health check pattern implementations in a sandbox.

Requirements

You need to install the following

• docker
• minikube
• bzt

Running the examples

Open the Makefile and change DOCKER and KUBECTL commands to according to your machine. To build the project and start the compose files you have to execute the following command:

make docker compose

You'll have Grafana with prebuilt dashboards on http://DOCKER_HOST:3000

To generate some load on the services, run the following:

make e2e

In the Makefile BASE_URL variable defines the endpoint for testing the application.

To start & stop containers you'll need to use the script in chaos-testing/kill-container.sh. It will look at containers flagged as killable and remove them from time-to-time.

Running with Kubernetes

make k8s-monitoring-deploy will deploy Prometheus in a monitoring namespace. make k8s-deploy will run the application with the dependencies under test namespace.

To generate some load on the services, run the following:

make e2e

You have to set the BASE_URL in the Makefile to te Kubernetes service endpoint first.

To start & stop pods you'll need to use the script in chaos-testing/whack-a-pod.sh. It will look at pods flagged as killable and remove them from time-to-time.

Interacting with the application

You can use the following endpoints for interacting with the deployed application.

HOST:PORT/application/hello - will return static response

HOST:PORT/application/mongo - will increment a counter in the deployed MongoDB database and change custom metrics

HOST:PORT/application/metrics - shows metrics endpoint

Cleanup

Remove deployments from test namespace and remove containers from compose files

make k8s-clean
make clean