Skip to content

Latest commit

聽

History

History
157 lines (101 loc) 路 6.92 KB

README.md

File metadata and controls

157 lines (101 loc) 路 6.92 KB
Raw

Chronos Kubernetes Example

This example demonstrates how Chronos can be used to track health metrics of a Kubernetes cluster when deployed with a Prometheus monitoring server. The folders in this example include the following:

  • client: A React application (in the client folder) that requests a random number from a Node/Express API
  • server: The Node/Express API that responds with the random number. This is where the Chronos package is actually imported and executed.
  • launch: YAML files describing the deployment and service configurations for the client, server, and Prometheus server
  • scripts: Scripts to automate some of the steps involved in running the example

Additional Documentation

For additional information on how Chronos works this example, please review the Kubernetes section in the Chronos NPM Package README.

Install Docker Desktop

This example has been developed and tested using the Kubernetes Engine packaged in the Docker Desktop application.

  1. Follow instructions online to download/install Docker Desktop and to enable the Kubernetes Engine.

enabled kubernetes engine in docker

Setup Prometheus and Grafana

  1. cd into the scripts folder and run the setup script with ./setup.sh - the process of configuring Prometheus and Grafana.

Note: If you run into Permission denied error, try run this command: chmod +x [the_file_name] in the terminal and re-run ./setup.sh.

  1. Open your web browser and navigate to localhost:32000. If you've correctly completed step 1 of the setup, you'll be directed to the Grafana login page. Use admin for both the username and password to log in. You can update the password after logging in if needed.

  2. Go to Home -> Administration -> Service accounts, and select Add service account to establish a new service account. Make sure to designate Admin as the role. Then, click on Add service account token, press generate, and you're finished! Keep this token in mind as you'll utilize it for accessing Grafana HTTP API programmatically.

  3. Add a .env file to the server folder that contains the following key/value pairs:

CHRONOS_GRAFANA_API_KEY = Bearer [the access token you created in step 3]

Note: If you are monitoring your own application, be sure to create a .env file as well.

Build the Client

  1. cd into the client folder and run the following command:
docker build -t frontend:1.0 .

Mac Users: Alternative to running the above command, cd into the scripts folder and run the buildClient.sh script by running ./buildClient.sh in the terminal.

Build the Server

  1. In the .env file you just created, add the following key/value pairs as well:
CHRONOS_DB = MongoDB or PostgreSQL
CHRONOS_URI = The URI to the desired MongoDB or PostgreSQL database to save health metrics via Chronos
  1. Then look at the package.json file in the server folder and note how @chronosmicro/tracker is included as a dependency:

  • If the @chronosmicro/tracker dependency is listed as a remote npm package (i.e. "@chronosmicro/tracker": "^12.0.1"), no further work is needed.

  • If the @chronosmicro/tracker dependency is listed as a local npm package (i.e. "@chronosmicro/tracker": "file:./chronos_npm_package"), the Docker build will require that the Chronos code is in this folder. Either:

    • Copy the chronos_npm_package folder in manually, and go to step 3
    • OR
    • If you are a Mac user, you can use the buildServer.sh script to automate this process and skip step 3
      • Do this by cding into the scripts folder and run the buildServer script with ./buildServer.sh - the process of copying in the chronos_npm_package folder, performing the Docker build, and then removing the copied in folder is automated.
  1. (skip if you used the buildServer.sh script)

cd into the server folder and run the following command:

docker build -t backend:1.0 .

Start the npm Server

  1. cd into the server folder inside chronos_npm_package, then run npm install to install all dependencies.

  2. run npm start to start npm server.

  3. You should see 馃 Example app listening on port 1111! chronos_npm_package Server Loaded 馃帀 in the terminal.

Deploy the Cluster

  1. cd into the launch folder and run the following commands to start the services and deployments described in the YAML files:
kubectl apply -f backend.yml
kubectl apply -f frontend.yml

Mac Users: Alternative to running the above commands, cd into the scripts folder and run the startKuber.sh script.

  1. Check in Docker desktop if your containers have been created. You should see something similar to the following:

Kubernetes containers created

dashboards created

The second image is inside k8s_node-backend_node-backend-d597768c-6zm5j_default_46da04f8-99c6-4522-9141-6c05f8d5141d_0 container, it may take a while to show up in Docker desktop based on the scraping interval.

Your microservice health metrics can now be viewed at the given CHRONOS_URI or, preferrably, in the Chronos desktop app.

Teardown the Cluster

  1. cd into the launch folder and run the following commands to stop the running services and deployments:
kubectl delete -f clusterRole.yml
kubectl delete -f promConfig.yml
kubectl delete -f prometheus.yml
kubectl delete -f backend.yml
kubectl delete -f frontend.yml

Mac Users: Alternative to running the above commands, cd into the scripts folder and run the stopKuber.sh script

Note: The above part only teardown Prometheus and Kubernetes, it leaves Grafana running. This is because if you teardown Grafana, the next time you redeploy you will be login with a new account, the access token and dashboard you created within this account will be lost.

To teardown grafana, run the following commands:

kubectl delete -f launch/grafana-datasource-config.yml
kubectl delete -f launch/grafanaService.yml
kubectl delete -f launch/grafana.yml

Insight for how to migrate this example to your own app

To configure Prometheus and Grafana, you can simply copy the below yaml files to your own app.

clusterRole.yml
promConfig.yml
prometheus.yml
grafana-datasource-config.yml
grafanaService.yml
grafana.yml

Then configure Prometheus and Grafana with your own application's deployments using images.

Contributing

Chronos hopes to inspire an active community of both users and developers. For questions, comments, or contributions, please submit a pull request.

Read our contributing README to further learn how you can take part in improving Chronos.

License

MIT