Demo setup¶

Follow these steps after Installation and when your cluster meets the Prerequisites. You will run a minimal httpbin workload behind ingress, attach an ElastiService, apply it, and curl through the ingress to see scale-from-zero and idle scale-down.

1. Deploy a Target Application¶

Before creating an ElastiService, you need a target deployment, service, and ingress for KubeElasti to manage. Below is a sample httpbin application you can use.

Create a file named target-deployment.yaml:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61

apiVersion: v1
kind: Namespace
metadata:
  name: target
---
apiVersion: v1
kind: Service
metadata:
  name: target-deployment
  namespace: target
spec:
  type: ClusterIP
  selector:
    app: target-deployment
  ports:
    - protocol: TCP
      port: 80
      targetPort: 80
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: target-deployment
  namespace: target
spec:
  replicas: 1
  selector:
    matchLabels:
      app: target-deployment
  template:
    metadata:
      labels:
        app: target-deployment
    spec:
      containers:
        - name: target-deployment
          image: kennethreitz/httpbin
          ports:
            - containerPort: 80
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: httpbin-ingress
  namespace: target
  annotations:
    nginx.ingress.kubernetes.io/rewrite-target: /$2
    nginx.ingress.kubernetes.io/service-upstream: "true"
    nginx.ingress.kubernetes.io/upstream-vhost: "target-deployment.target.svc.cluster.local"
spec:
  ingressClassName: nginx
  rules:
    - http:
        paths:
          - path: /httpbin(/|$)(.*)
            pathType: ImplementationSpecific
            backend:
              service:
                name: target-deployment
                port:
                  number: 80

The nginx.ingress.kubernetes.io/upstream-vhost annotation above is important for elasti: it sets the Host header on the request NGINX forwards upstream, and the resolver reads that header to decide which target service to route to. See Resolver Architecture > Request Routing for details on how routing works and how to override the header.

Apply it:

kubectl apply -f target-deployment.yaml

Verify the target is running:

kubectl get pods -n target

2. Define an ElastiService¶

You turn scale-to-zero on for a workload by creating an ElastiService object: it points at your Kubernetes Service and scale target, defines Prometheus triggers, and (optionally) links a KEDA ScaledObject so KubeElasti can pause it while the workload is at zero. For a full walkthrough of every field and option, read the Configuration guide.

The manifest below matches the httpbin example from the previous step. Save it as elasti-service.yaml; you will apply it in the next section.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

apiVersion: elasti.truefoundry.com/v1alpha1
kind: ElastiService
metadata:
  name: target-elastiservice
  namespace: target
spec:
  service: target-deployment
  minTargetReplicas: 1
  cooldownPeriod: 5
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: target-deployment
  triggers:
    - type: prometheus
      metadata:
        query: sum(rate(nginx_ingress_controller_requests{namespace="target", ingress="httpbin-ingress"}[1m])) or vector(0)
        serverAddress: http://kube-prometheus-stack-prometheus.monitoring.svc.cluster.local:9090
        threshold: "0.01"
  probeResponse:
    - method: GET
      path:
        type: PathPrefix
        value: /healthz
      response:
        status: 200
        body: '{"ok":true}'

If your ingress, load balancer, or platform sends health checks while the workload is at zero replicas, add probeResponse rules for those paths. Requests that match are served by the resolver and do not scale the deployment up.

3. Apply the KubeElasti service configuration¶

Apply the configuration to your Kubernetes cluster:

kubectl apply -f elasti-service.yaml -n <service-namespace>

The pod will be scaled down to 0 replicas if there is no traffic.

4. Test the setup¶

You can test the setup by sending requests to the nginx load balancer service.

# For NGINX
kubectl port-forward svc/nginx-ingress-ingress-nginx-controller -n nginx 8080:80

# For Istio
kubectl port-forward svc/istio-ingressgateway -n istio-system 8080:80

Start a watch on the target deployment.

kubectl get pods -n <NAMESPACE> -w

Send a request to the service.

curl -v http://localhost:8080/httpbin

You should see the pods being created and scaled up to 1 replica. A response from the target service should be visible for the curl command. The target service should be scaled down to 0 replicas if there is no traffic for cooldownPeriod seconds.