Deploy Umami on CCE

In this blueprint we are going to set up Umami on Open Telekom Cloud's Cloud Container Engine (CCE), leveraging Kubernetes for scalability and flexibility. For the database backend, we will use the Zalando PostgreSQL Operator to provision and manage a PostgreSQL cluster within the CCE environment.

Prerequisites

We are going to need a CCE Cluster (its provisioning is out of the scope of this blueprint) and the zalando-postgres-operator. This operator automates the management of PostgreSQL clusters on Kubernetes, handling tasks like scaling, backups, and failover. It simplifies the deployment and maintenance of a highly available PostgreSQL database within the CCE cluster. Additionally we are going to need an Elastic Load Balancer in order to expose Umami.

Installing Zalando Postgres Operator

We are going to install the operator by using the provided Helm chart:

helm repo add postgres-operator-charts https://opensource.zalando.com/postgres-operator/charts/postgres-operator
helm repo update

helm install postgres-operator postgres-operator-charts/postgres-operator

Installing Umami

Provisioning a Database

As we priorly discussed, we are going to use zalando-postgres-operator in order to provision a PostgreSQL Cluster in CCE:

umami-postgresql.yaml

apiVersion: acid.zalan.do/v1
kind: postgresql
metadata:
  labels:
    application: umami
  name: umami-psql
spec:
  databases:
    umami: umami
  numberOfInstances: 1
  postgresql:
    version: "16"
    parameters:  
      huge_pages: "false"
  preparedDatabases:
    umami:
      defaultUsers: true
      schemas:
        data: {}
        history:
          defaultRoles: true
          defaultUsers: false
  resources:
    limits:
      cpu: 500m
      memory: 500Mi
    requests:
      cpu: 10m
      memory: 100Mi
  teamId: default
  users:
    admin:
      - superuser
      - createdb
    umami: []
  volume:
    size: 1Gi
    storageClass: csi-disk

kubectl apply -f umami-postgresql.yaml

Deploying Umami

Create the follow manifest:

umami-web-deployment.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: umami-web    
spec:
  replicas: 2
  selector:
    matchLabels:
      app: umami-web
  template:
    metadata:
      labels:
        app: umami-web
    spec:
      containers:
        - name: web
          image: ghcr.io/umami-software/umami:postgresql-latest
          ports:
            - containerPort: 5000
              protocol: TCP
          env:
            - name: PORT
              value: '5000'
            - name: DB_DATABASE
              value: "umami"
            - name: DB_HOST
              value: umami-psql.docs-next.svc.cluster.local
            - name: DB_PORT
              value: '5432'
            - name: DB_USERNAME
              valueFrom:
                secretKeyRef:
                  name: umami.umami-psql.credentials.postgresql.acid.zalan.do
                  key: username
            - name: DB_PASSWORD
              valueFrom:
                secretKeyRef:
                  name: umami.umami-psql.credentials.postgresql.acid.zalan.do
                  key: password
            - name: DATABASE_URL
              value: "postgres://$(DB_USERNAME):$(DB_PASSWORD)@$(DB_HOST):$(DB_PORT)/$(DB_DATABASE)"
          imagePullPolicy: IfNotPresent

kubectl apply -f umami-web-deployment.yaml

important

A Kubernetes Secret with the name umami.umami-psql.credentials.postgresql.acid.zalan.do, containing the credentials of the umami database, will be automatically provisioned by the zalando-postgres-operator during the application of manifest umami-postgresql.yaml. The environmental variables DB_USERNAME & DB_PASSWORD are getting their values by referencing this Secret.

Creating an Elastic Load Balancer

Navigate to Network Console->Elastic Load Balancing and click Create Elastic Load Balancer. Choose to create Shared Load Balancer and choose New EIP so the new ELB is automatically bound to a new elastic IP:

tip

Write down the ID of the Elastic Load Balancer we are going to need it in the next steps.

Exposing Umami

Creating a Service

umami-service.yaml

apiVersion: v1
kind: Service
metadata:
  name: umami-web
spec:
  type: NodePort
  ports:
    - protocol: TCP
      name: umami
      port: 5000
      targetPort: 5000
  selector:
    app: umami-web

kubectl apply -f umami-service.yaml

note

If you are not planning to expose the service via an Ingress object, change the type to ClusterIP.

Creating an Ingress (optional)

umami-ingress.yaml

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: umami-ingress
  labels:
    app: umami-web
  annotations:
    kubernetes.io/elb.class: union
    kubernetes.io/elb.id: {value}
    kubernetes.io/elb.port: 80
spec:
  ingressClassName: cce
  rules:
  - host: umami.example.com
    http:
      paths:
      - backend:
          service:
            name: umami-web
            port:
              number: 5000
        path: /
        pathType: ImplementationSpecific

important

• Replace the placeholder {value} of annotation kubernetes.io/elb.id with the ID of the Elastic Load Balancer we created before.
• If the Elastic Load Balancer you created was a shared one then the annotation kubernetes.io/elb.class should have the value union.
• Replace umami.example.com in host, with the FQDN of yours.

kubectl apply -f umami-ingress.yaml

Verification

Open in a browser the address: http://ELB_EIP and you should now land at the logon page of Umami:

warning

Umami uses admin/umami as default credentials. Change them immediatelly after you log in!