24. Day 24 -- Ingress, Metrics Server & Horizantol Pod Auto Scaling.md

Day 24 -- Ingress, Metrics Server & Horizantol Pod Auto Scaling

Ingress

Ingress Exposes multiple Services through a single cloud load-balancer.

Ingress vs Load-balancer.png

Ingress Fanout

Metrics Server

Collects resource metrics from kubelets & exposes them in K8s API server through metrics API for use by HPA & VPA

Horizontal Pod Auto Scaling

HorizantolPodAutoScaler automatically updates workload resources to match demand.

Horizantol Scaling --> Deploy more Pods.
Vertical Scaling --> Assign more resources (CPU, memory, etc) to existing pods.

HPA Controller

Adjusts the desired scale of the workloads to match observed metrics (average CPU utilization, etc)

desiredReplicas = ceil[currentReplicas * (currentMetricValue / desiredMetricValue)]