grpc-lab

grpc-lab

Simple gRPC service with HTTP extension based on a proto API definition. Includes deployment to Kubernetes and ingress configuration with TLS certificate and custom domain.

setup

Export a couple of environment variables to make this readme reproducible:

Choose domain for which you can control. When setting up ingress, you will have to create DNS entry.

export DOMAIN="thingz.io"
export SPACE="grpc-lab"

Create namespace:

kubectl create ns $SPACE

deployment

All commands assume you are executing from the root of the repo

Apply deployment:

kubectl apply -f deployments/app.yaml -n $SPACE
kubectl rollout status deployment.apps/ping -n $SPACE

Check pod status:

kubectl get pods -n $SPACE

Response should look something like this:

NAME                    READY   STATUS    RESTARTS   AGE
ping-554f558fbd-h96rf   1/1     Running   0          15s

You can also check server logs:

kubectl logs -l app=ping -n $SPACE 

service

kubectl apply -f deployments/service.yaml -n $SPACE

kubectl get service -n $SPACE

Response should look something like this:

NAME   TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)     AGE
ping   ClusterIP   10.0.31.13   <none>        50505/TCP   6s

ingress

Create TLS certificates:

certbot certonly --manual \
                 --preferred-challenges dns \
                 -d "*.${DOMAIN}" \
                 --config-dir ./certs \
                 --logs-dir ./certs \
                 --work-dir ./certs

When promoted to configure DNS TXT record, you can check on the propagation status by digging dig "_acme-challenge.${DOMAIN}" TXT

Once the certs were generated, create the TLS secret:

kubectl create secret tls tls-secret \
		--key "live/${DOMAIN}/privkey.pem" \
		--cert "live/${DOMAIN}/fullchain.pem" \
		-n $SPACE 

Create gRPC and HTTP ingresses

kubectl apply -f deployments/grpc-ingress.yaml -n $SPACE
kubectl apply -f deployments/http-ingress.yaml -n $SPACE

Watch the ingresses until the ADDRESS column gets populated:

kubectl get ingress -n $SPACE -w

Now go to the DNS server and create an A entry for the * and the IP value in the HOSTS column. You can test when this gets propagated using dig gping.thingz.io for example, or whatever the HOST value is.

test

Once the DNS is set up, test the gRPC endpoint:

grpcurl -d '{"id":"id1", "message":"hello"}' \
  gping.thingz.io:443 \
  io.thingz.grpc.v1.Service/Ping

Response should look something like this:

{
  "id": "id1",
  "message": "hello",
  "reversed": "olleh",
  "count": "1",
  "created": "1606055127674975938",
  "metadata": {
    "address": "[::]:50505"
  }
}

Now, test the HTTP endpoint:

curl -i -d '{"id":"id1", "message":"hello"}'\
      -H "Content-type: application/json" \
      https://ping.thingz.io:443/v1/ping

Again, the response should look something like this:

HTTP/2 200
date: Sun, 22 Nov 2020 14:27:20 GMT
content-type: application/json
content-length: 129
grpc-metadata-content-type: application/grpc
strict-transport-security: max-age=15724800; includeSubDomains

{
  "id":"id1",
  "message":"hello",
  "reversed":"olleh",
  "count":"4",
  "created":"1606055240599449988",
  "metadata":{
    "address":"[::]:50505"
  }
}

cleanup

kubectl delete -f deployments/ -n $SPACE
kubectl delete secret tls-secret -n $SPACE 
kubectl delete ns $SPACE