Fifty Shades of Kubernetes Autoscaling

Is your Kubernetes autoscaling just HPA? Discover the layers of elasticity that slash costs and handle any demand.

#1about 4 minutes

Why cloud-native systems require multi-layered elasticity

Modern applications need to be anti-fragile and support hyperscale, which requires elasticity at the workload level (horizontal/vertical) and the infrastructure level (cluster scaling).

#2about 5 minutes

How metrics and events drive Kubernetes autoscaling decisions

Autoscaling relies on events for cluster-level actions and a multi-layered metrics API for workload scaling based on resource, custom, or external data sources.

#3about 5 minutes

Implementing horizontal pod autoscaling with different metrics

The Horizontal Pod Autoscaler (HPA) can scale pods based on simple resource metrics like CPU, custom pod metrics, or external metrics from Prometheus.

#4about 2 minutes

Using the vertical pod autoscaler for right-sizing workloads

The Vertical Pod Autoscaler (VPA) can automatically adjust pod resources, but its recommendation mode is most useful for determining optimal CPU and memory settings.

#5about 4 minutes

How the default cluster autoscaler works on GKE

The default cluster autoscaler automatically provisions new nodes when it detects unschedulable pods due to resource constraints, as demonstrated on Google Kubernetes Engine.

#6about 5 minutes

Using Carpenter for fast and flexible cluster scaling on AWS

Carpenter provides a fast and flexible cluster autoscaling solution for AWS EKS, enabling cost optimization by using spot instances for scaled-out nodes.

#7about 1 minute

Exploring KEDA for advanced event-driven autoscaling

KEDA (Kubernetes Event-driven Autoscaling) enables scaling workloads, including to zero, based on events from various sources like message queues or databases.

#8about 1 minute

Summary of Kubernetes autoscaling tools and techniques

A recap of essential autoscaling components including the metric server, HPA, VPA, cluster autoscalers like Carpenter, KEDA, and the descheduler for cluster optimization.

#9about 2 minutes

Q&A on autoscaler reliability and graceful shutdown

Discussion on the production-readiness of autoscalers, the importance of observability, and how to achieve graceful pod termination during scale-down events.