The demo-apps-project deploys various web applications onto Kubernetes / GCP using the Google Kubernetes Engine (GKE).
By default, this project ships with minimally configured networking for simplicity and ease of setup, though our services will be available only privately via kubectl proxy
. For this project, you can enable the proxy by typing in your shell
docker-compose up
Use can then access the applications using this format:
http://localhost:8001/api/v1/namespaces/<NAMESPACE>/services/<SERVICE-NAME>:<SERVICE-PORT>/proxy/
If we want to expose our services (applications) on the public Internet, we will need to:
- Create an external static IP address.
- Enable
nginx-ingress
module (to create Kubernetes LoadBalancer type of service) that will use that static IP address. - Create DNS records to point clients to the static IP address.
- Enable cert-manager module to add TLS for the domains
- Enable
Ingress
resources for our Helm releases.
Use this pull request as an example of how to set up production networking.
(TODO) There is also a step-by-step tutorial: https://docs.exekube.com/in-practice/production-networking
This project uses the Exekube base-project as boilerplate.
Modules from base-project:
- gke-network
- gke-cluster
- administration-tasks
We use nginx-ingress as our ingress controller for the applications.
We use cert-manager to acquire Let's Encrypt TLS certificates to enable HTTPS for publicly available appications.
Guestbook is a canonical example of a Kubernetes deployment
- Private address: http://localhost:8001/api/v1/namespaces/default/services/frontend:80/proxy/
- Public address: read instructions above
A custom React app to test out Istio ingress and auto-injection
- Private address: http://localhost:8001/api/v1/namespaces/default/services/forms-app-nginx-react:80/proxy/
- Public address: read instructions above