KEDA 2.11: Expanded Autoscaling and Enhanced Scalers

Introduction

KEDA 2.11, released on September 5, 2024, continues to expand the capabilities of event-driven autoscaling on Kubernetes. This release introduces new scalers, performance improvements, enhanced security features, and better integration with Kubernetes Horizontal Pod Autoscaler (HPA).

New Scalers

• Redis Streams scaler enables autoscaling based on Redis stream message backlog.
• Apache Pulsar scaler provides autoscaling for Pulsar message queue workloads.
• MongoDB scaler enables autoscaling based on MongoDB collection metrics.
• Custom metrics scaler allows integration with any metrics provider through a flexible API.

Performance Improvements

1. Scaler evaluation optimizations reduce CPU usage during metric collection.
2. Caching enhancements improve response times for frequently accessed metrics.
3. Concurrent scaling improvements enable better handling of multiple scaled objects.
4. Resource usage optimizations reduce memory footprint in large clusters.

Security Enhancements

• Authentication improvements support more authentication methods for scalers.
• Secret management integration with external secret operators provides secure credential handling.
• RBAC refinements provide more granular permissions for KEDA components.
• Audit logging tracks all scaling decisions and metric evaluations for compliance.

HPA Integration

• Metrics API improvements provide better integration with Kubernetes metrics API.
• External metrics support enables scaling based on custom external metrics.
• Behavior configuration allows fine-grained control over scaling behavior.
• Stabilization windows provide smoother scaling decisions with reduced oscillations.

Getting Started

# Install KEDA
kubectl apply -f https://github.com/kedacore/keda/releases/download/v2.11.0/keda-2.11.0.yaml

Create a ScaledObject for Redis:

apiVersion: keda.sh/v1alpha1
kind: ScaledObject
metadata:
  name: redis-scaledobject
spec:
  scaleTargetRef:
    name: my-deployment
  minReplicaCount: 1
  maxReplicaCount: 10
  triggers:
  - type: redis
    metadata:
      address: redis-service:6379
      listName: mylist
      listLength: "10"

Summary

KEDA 2.11 continues to make Kubernetes the ideal platform for event-driven workloads with flexible and efficient autoscaling.