ingress-nginx-app release v1.6.0

Changed

• Upgrade to nginx-ingress-controller 0.30.0. (#31)
• Configured app icon. (#32)
• Enabled HorizontalPodAutoscaler by default. (#27)
• Based on HPA trials done so far, following settings have been adjusted to better fit actual observed usage profiles:
  • CPU resource requests have been adjusted from 500m to 2 CPU
    • 0.5 CPU was not enough for all the processes NGINX Ingress Controller starts
  • Memory requests changed from 600Mi to 2.5GB
    • Scaling out does not shard Ingress definitions and other configurations stored in memory of every nginx-ingress-controller replica
    • Memory usage spikes during configuration reloads
    • It improves the HPA stability
  • Default number of nginx worker processes was changed from 4 to 1
    • This reduced memory usage of each replica
    • It didn’t affect request handling capacity
    • Better defaults considering CPU requests and number of processes running on every nginx-ingress-controller replica.
• To avoid cluster-operator and HPA collision and nginx service disruption, this release also breaks with cluster-operator controllable nginx ingress controller Deployment replicas count
  • ingressController.replicas which was previously dynamically set by cluster-operator is now removed
  • New controller.replicaCount config property is introduced, default replica count is set to 1, and then by default enabled HPA takes it over from there, by default scaling the Deployment in range of 1 to 20 replicas
  • If HPA gets disabled on-demand, replica count will stay static if not manually or automatically changed by some third party.