Redeploy Prometheus

Before starting

• Start docker Desktop as admin
• Ensure that kubernetes is running in docker desktop if it's not starting, go in the debug menu and click on clean/purge data
• Start Lens as admin
• Open powershell as admin
• Pull latest version of k8s_mathematech

Clean

• In lens in the deployment menu, kill all deployments without kube-system namespace
• In lens in the pods menu, kill all services without the kube-system namespace
• Same for network policies, ports expose, etc.

Redeploy in local

Be carefull because you will not probably have python in your path, because you are logged in the powershell as admin. So you can run python by paste the absolute path of the python executable.

git fetch --prune
python -m deploy -u -i database:*.* -c local
python -m deploy -u -i ingress-nginx:*.* -c local
python -m deploy -u -i metrics:*.* -c local

Access to Grafana

Do not expose port to access to grafana, nginx will do it for you. http://localhost/apps/grafana/

In production but you probably don't want to do that

git fetch --prune
python3 -m deploy -u -i database:postgresql -c local
python3 -m deploy -u -i database:postgresql -c prod
git pull  && python3 -m deploy -u -i metrics:deadline-events -c prod
cd k8s_mathematech/
git pull  && python3 -m deploy -u -i metrics:deadline-events -c prod
cd k8s_mathematech/
python3 -m deploy -u -i metrics:grafana -c prod

Add Metric Service

Edit the file k8s_mathematech\configurations\metricsprometheus-license-exporter for example.

You will:

• Configmap
• Deployment
• Service
• Local\servicd

Configmap

Configmap are used to store configuration in kubernetes. It's a key value store. So for example if you wan to add a new service for prometheus you can do this:

apiVersion: v1
kind: ConfigMap
metadata:
  name: redshift-prometheus-rlm-exporter-configmap
  namespace: metrics
data:
  RLM_SERVER_HOST: "lic-02.mtc.wtf"
  RLM_SERVER_PORT: "5053"
  RLM_SERVER_USERNAME: ""
  RLM_SERVER_PASSWORD: ""
  PORT: "8082"

Deployment

Here you will describe the deployment of the service. You will describe the image, the ports, the environment variables, etc.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: redshift-prometheus-rlm-exporter
  namespace: metrics
  labels:
    app: redshift-prometheus-rlm-exporter
spec:
    selector:
      matchLabels:
        # This will be our deployment label
        app: prometheus-license-exporter # this is really important, 
  template:
    metadata:
      labels:
        app: prometheus-license-exporter
    spec:
      containers:
      - name: prometheus-license-exporter
        resources:
          requests:
            memory: "64Mi"
            cpu: "250m"
          limits:
            memory: "128Mi"
            cpu: "500m"
        image: registr.mtc.wtf/prometheus-license-exporter:latest
        imagePullPolicy: Always
        # if you are in dev:
        # image: yourlocalimagetag
        # imagePullPolicy: Never    
        envFrom:
            - configMapRef:
                name: redshift-prometheus-rlm-exporter-configmap
          ports:
            - containerPort: 8082
      volumes:
        - name: license-config
          configMap:
            name: prometheus-license-exporter-configmap

Service

You will also fin a service.yml file under local folder, this is for the -c local using python -m deploy -u -i metrics:.* -c local command.

Service.ymll will describe the service, the port, the target port, etc.

...
---
apiVersion: v1
kind: Service
metadata:
  name: redshift-prometheus-rlm-exporter
  namespace: metrics
  labels:
    app: prometheus-license-exporter # this is really important and should match the deployment
  annotations:
    prometheus.io/path: /metrics
    prometheus.io/port: "8082"
    prometheus.io/scrape: "true"
spec:
  selector:
    app: prometheus-license-exporter # this is really important and should match the deployment
  ports:
    - protocol: TCP
      port: 8082
      targetPort: 8082
      name: http