k8s/monitoring/README.md

-# Deploy the Prometheus operator on minikube
+# Monitoring hands-on
 
-## Install the prometheus stack
+## Setup DNS names
 
+Append the following line on the /etc/hosts on your machine (your laptop):
+
+```
+<your-olss-vm-ip> minikube grafana.minikube prometheus.minikube olss-demo-app.minikube ingress-example.minikube
+```
+
+This is what I have on my mac:
+```
+❯ tail -1 /etc/hosts
+131.154.97.229 minikube grafana.minikube prometheus.minikube olss-demo-app.minikube tutor-01.olss ingress-example.minikube
+```
+
+If you want to also contact the minikube ingress from within the OLSS VM,
+login into the OLSS VM, collect the minikube ip with:
+
+```
+minikube ip
+```
+
+and append the following on the /etc/hosts on that machine, e.g.:
+
+```
+<your-minikube-ip> minikube grafana.minikube prometheus.minikube olss-demo-app.minikube ingress-example.minikube
+```
+
+On my tutor-1 machine I have:
+
+```
+[centos@tutor1 ~]$ minikube ip
+192.168.49.2
+[centos@tutor1 ~]$ tail -1 /etc/hosts
+192.168.49.2 minikube grafana.minikube prometheus.minikube olss-demo-app.minikube ingress-example.minikube
+```
+
+Make sure the minikube ingress add is enabled:
+```
+[centos@tutor1 monitoring]$ minikube addons list | grep ingress
+| ingress                     | minikube | enabled ✅   |
+```
+
+If is not enabled, enable it:
+```
+minikube addons enable ingress
+```
+
+# Install the OLSS demo app in your cluster
+
+```
+git clone https://github.com/andreaceccanti/olss-demo-app/
 ```
+
+If have already cloned in the past days, or made your own fork, clone it 
+again in another repo and use that one:
+
+```
+git clone https://github.com/andreaceccanti/olss-demo-app/
+olss-demo-app-orig
+```
+
+Go to the `k8s/olss-demo-app` folder and type the following command:
+
+```
+kubectl apply -k .
+```
+
+This command will install the demo app on your cluster.
+
+```
+[centos@tutor1 olss-demo-app]$ kubectl -n olss-demo-app get po
+NAME                            READY   STATUS    RESTARTS   AGE
+olss-demo-app-58cd985fc-8drr8   1/1     Running   0          58s
+```
+
+# Monitoring probes
+
+https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-startup-probes/
+
+Deploy the OLSS demo app without probes:
+
+```
+kubectl apply -f olss-app.deploy.no-probes.yaml
+```
+
+# Create a kube config context for the olss-demo-app namespace
+
+```
+kubectl config set-context --cluster=minikube --user=minikube --namespace=olss-demo-app olss-demo-app
+kubectl config use-context olss-demo-app
+```
+
+# Install the multitool container
+
+```
+kubectl run multitool --image=praqma/network-multitool --restart Never
+```
+
+Enter the multitool container:
+
+```
+kubectl exec -ti multitool -- sh
+```
+
+Within the mulitool container run the following command:
+
+```
+watch -n.5 curl olss-demo-app:8080/api/hello
+```
+
+The above commands issue 2 requests per second against the OLSS demo app.
+
+Now, in another terminal, scale up the olss-demo-app deployment to 3 replicas:
+
+```
+kubectl scale deployment --replicas=2 olss-demo-app
+```
+
+You will observe that while containers start up some requests fail.
+This is because we do not have probes enabled.
+
+Scale down the deployment back to 1 replica:
+
+```
+kubectl scale deployment --replicas=1 olss-demo-app
+```
+
+Deploy the olss-demo-app with probes:
+
+```
+kubectl apply -f olss-app.deploy.yaml
+
+```
+
+Scale up the deployment again. This time you will not see errors, since the
+probes configure the request routing only when the replica is in good shape to
+serve it.
+
+# Install prometheus on your cluster 
+
+
+```
+# Install prometheus community helm charts
+helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
+
+# create a namespace for prometheus and monitoring-related stuff
 kubectl create namespace monitoring
+
+# install the prometheus stack in the monitoring namespace
 helm install prometheus prometheus-community/kube-prometheus-stack --namespace monitoring --values=prometheus-kube-stack.values.yml
 ```
+
+# Setup port forwarding to reach the ingress controller
+
+In another terminal running on the OLSS VM, type the following command
+(and leave it running):
+
+
+```
+kubectl --namespace=ingress-nginx port-forward --address 0.0.0.0
+svc/ingress-nginx-controller 8080:80
+```
+
+
+# Open the grafana dashboard in your browser
+
+Point your browser to:
+
+http://grafana.minikube:8080
+
+You should see the Grafana login page.
+
+You can log into the system with the following credentials:
+
+- username: admin
+- password: prom-operator
+
+# Open prometheus dashboard in your browser
+
+Point your browser to:
+
+http://promethes.minikube:8080
+
+You should see the Prometheus dashboard.
+
+# Setup monitoring for the olss-demo-app
+
+Go back to the olss-demo-app repo you cloned, cd into `k8s/olss-demo-app` and
+run the following command:
+
+```
+kubectl apply -f olss-app.servicemonitor.yaml
+```
+
+In a while prometheus should start scraping metrics from the demo app.
+

pom.xml

@@ -10,7 +10,7 @@
   </parent>
   <groupId>it.infn</groupId>
   <artifactId>olss-demo-app</artifactId>
-  <version>0.0.4</version>
+  <version>0.0.5</version>
   <name>olss-demo-app</name>
   <description>Demo project</description>