RPaaS v2 (short for Reverse Proxy as a Service v2) is a system that allows managing fully configurable Nginx web servers as simple Tsuru service instances. With it, developers can expose their apps on Tsuru to the wild using the old fashion Tsuru API and tooling.
Powered by Nginx OSS, developers can use several features such as:
- Advanced L7 routing;
- Cache;
- Canary deployments of Tsuru apps;
- Web Application Firewall - WAF;
- Rate limit / DDoS protection;
- Golden signals;
- GeoIP;
- TLS termination (w/ session resumption);
- and so on.
NOTE: As you may realize this project it's pretty similar to Nginx Ingress but each RPaaS instance provides an fully independent Nginx web server.
This project consists of several parts, namely:
- RPaaS v2 API: web API which implements the Tsuru service contract.
- RPaaS v2 purger API: web API which allows purging the cached resources in live Nginx pods;
- RPaaS v2 plugin: command line tool to manage RPaaS instances (compatible with Tsuru plugin - aka
tsuru rpaasv2
); - Kubernetes operator: CRD and controller which allows managing RPaaS instances using the Kubernetes API.
You will need a local Kubernetes cluster. There are bunch of tools to provision it, we recommend kind.
Warning: Keep in mind that all following steps are going to use the Kubernetes cluster from your current KUBECONFIG
config file (which defaults to ~/.kube/config
).
After cluster created, you need to install the Nginx Operator and Cert Manager with Helm.
# installing Nginx Operator
helm repo add tsuru https://tsuru.github.io/charts
helm install --create-namespace --namespace nginx-operator nginx-operator tsuru/nginx-operator
# installing Cert Manager
helm repo add jetstack https://charts.jetstack.io
helm install --create-namespace --namespace cert-manager --set installCRDs=true cert-manager jetstack/cert-manager
In a terminal session, run the following commands to start the Kubernetes operator subsystem.
# generate and apply all CRDs to Kubernetes
make deploy
# running the controller
make run
In another terminal session, run the following command to start the RPaaS v2 API.
# starts the web API on :9999/TCP
make run/api
After following the steps above you can create and manage the instance from RPaaS v2 plugin:
# creates a RPaaS instance
kubectl apply -n rpaasv2 -f test/testdata/rpaas-full.yaml
# builds and uses the RPaaS v2 plugin to fetch infos about the instance created above
make build/plugin/rpaasv2
./bin/rpaasv2 --rpaas-url http://localhost:9999 info -s rpaasv2 -i my-instance
If you have the tsuru-client installed, it is as easy as running:
tsuru plugin install rpaasv2 "https://github.com/tsuru/rpaas-operator/releases/latest/download/manifest.json"