Horizontal Pod Autoscaler (HPA)
Automatically scale your deployments up and down based on CPU, memory, or custom metrics using the Horizontal Pod Autoscaler.
Full content coming soon. This lesson is being written.
This lesson covers HPA:
- HPA architecture: Metrics Server, HPA controller, target deployment
- Scaling on CPU utilization: the most common pattern
- Scaling on memory (with caveats)
- Custom metrics: KEDA, Prometheus Adapter
- Scale-down cooldown:
--horizontal-pod-autoscaler-downscale-stabilization - Min and max replicas: guarding against over/under-scaling
- Conflict with manual scaling
- HPA v2 API: multiple metrics, external metrics
Check back soon for the full content.