NGINX Ingress¶

Use this guide when NGINX Ingress Controller fronts workloads managed by KubeElasti.

Resolver header¶

Leave headerForHost at Host (the resolver binary default). If you installed via the Helm chart with its default X-Envoy-Decorator-Operation, override:

helm upgrade --install elasti ./charts/elasti -n elasti \
  --set elastiResolver.proxy.env.headerForHost=Host

Install NGINX Ingress¶

helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx
helm repo update
helm upgrade --install nginx-ingress ingress-nginx/ingress-nginx \
  --namespace nginx \
  --set controller.metrics.enabled=true \
  --set controller.metrics.serviceMonitor.enabled=true \
  --create-namespace

Enable Prometheus as described in Prerequisites so NGINX metrics are available for ElastiService triggers.

Route traffic to the target¶

The annotation nginx.ingress.kubernetes.io/upstream-vhost sets the upstream Host header to the service FQDN before the request reaches the resolver. The resolver reads that header to select the target service.

Example Ingress for the demo setup httpbin workload (target-deployment in namespace target):

httpbin-ingress.yaml
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

service-upstream: "true" sends traffic to the cluster Service (including the Elasti-managed endpoint) rather than bypassing it via Endpoints.

Prometheus trigger¶

Example query for scale-down when ingress traffic is idle:

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"

More examples are in Triggers.

Test (port-forward)¶

kubectl port-forward svc/nginx-ingress-ingress-nginx-controller -n nginx 8080:80
curl -v http://localhost:8080/httpbin