KEDA
Event-driven autoscaling for Kubernetes workloads.
About
KEDA extends Kubernetes autoscaling so workloads can scale from events instead of only CPU and memory. It connects Deployments and Jobs to scalers for Kafka, RabbitMQ, AWS SQS, Azure Service Bus, Redis, Prometheus, cron schedules, and many other sources, including scale-to-zero for idle event-driven workloads.
Best for
Teams running queue consumers, stream processors, scheduled jobs, and other event-driven workloads that should scale based on real work waiting outside the cluster.
★★★★☆ 4.6/5
Pros
- +Supports event-driven scaling from dozens of external systems
- +Enables scale-to-zero for async workers and batch-style workloads
- +Works with Kubernetes HPA instead of replacing core autoscaling primitives
- +CNCF graduated project with strong production adoption
Cons
- -Scaler configuration and credentials need careful namespace management
- -Cold starts can matter for workloads that scale from zero
- -Polling intervals and max replica counts need tuning to avoid surprise load or cost
Featured badge
Add this badge to your website to show that KEDA is featured on KubeThings.
Dark badge
<script async src="https://kubethings.com/embed/featured.js" data-kubethings-tool="keda"></script>Light badge
<script async src="https://kubethings.com/embed/featured.js" data-kubethings-tool="keda" data-kubethings-theme="light"></script>Comments
Comments are moderated before they appear.
No approved comments yet. Start the discussion.
Related Tools
All tools ->
Karpenter
Open source Kubernetes node autoscaling and just-in-time provisioning.
Kubernetes / AWS / Cloud Native