What is Keda in the context of Autoscaling?

What is Keda in the context of Autoscaling?

• A. It is a monitoring tool for cluster health
• B. It is a cloud service for storing metrics
• C. It is a performance tuning tool for autoscalers
• D. It is an event driven architecture that makes use of ScaledObjects

Answer: (D)

KEDA is an event-driven autoscaler for Kubernetes that adjusts pod counts based on external triggers, rather than just in-cluster metrics. It integrates with the standard Horizontal Pod Autoscaler by exposing scaling decisions through a ScaledObject, a Kubernetes Custom Resource that ties a workload (like a Deployment) to one or more triggers (such as queue length, messages in a pub/sub system, or external metrics from Prometheus). The ScaledObject specifies min and max replicas and the triggers that define how scaling should occur. This lets you scale up when there’s activity in an external system and scale down to zero when there isn’t, without modifying the application code. In short, KEDA is built for event-driven scaling and uses ScaledObjects to declare how to scale based on external events.

KEDA is an event-driven autoscaler for Kubernetes that adjusts pod counts based on external triggers, rather than just in-cluster metrics. It integrates with the standard Horizontal Pod Autoscaler by exposing scaling decisions through a ScaledObject, a Kubernetes Custom Resource that ties a workload (like a Deployment) to one or more triggers (such as queue length, messages in a pub/sub system, or external metrics from Prometheus). The ScaledObject specifies min and max replicas and the triggers that define how scaling should occur. This lets you scale up when there’s activity in an external system and scale down to zero when there isn’t, without modifying the application code. In short, KEDA is built for event-driven scaling and uses ScaledObjects to declare how to scale based on external events.