All tasks
Jump to a procedure below, or browse Scaling, Operate and Best Practices.
Getting started
Section titled “Getting started”Install
Section titled “Install”- Install with Helm
- Install with Argo CD
- Install with Terraform
- Install from a private registry
- Migrate an existing KEDA installation
- Install and verify multi-tenant scaling
Install: Cluster setup
Section titled “Install: Cluster setup”Install: Marketplaces & catalogs
Section titled “Install: Marketplaces & catalogs”Scaling: OpenTelemetry: Guides
Section titled “Scaling: OpenTelemetry: Guides”- Scale a workload from an OTel metric
- Migrate scaling metrics from Prometheus to OTel
- Scale from Ingress NGINX metrics with OTel
- Scale vLLM with OTel model metrics
Scaling: HTTP: Routing integrations
Section titled “Scaling: HTTP: Routing integrations”- In-Cluster HTTP Autoscaling
- HTTP Scaling for Ingress-Based Applications
- HTTP Scaling with OpenShift Routes
- HTTP Scaling with Kubernetes Gateway API
- HTTP Scaling with Istio VirtualServices
- HTTP Scaling with Gloo Gateway
- TLS Ingress HTTP Autoscaling
- HTTP Scaling with Argo Rollouts Canary
- Scale inference behind an Ingress
Scaling: HTTP: Guides
Section titled “Scaling: HTTP: Guides”- HTTP Scaler Waiting & Maintenance Pages
- Trace HTTP requests through Kedify
- Tune Kedify Proxy performance
Scaling: Vertical: Guides
Section titled “Scaling: Vertical: Guides”Insights
Section titled “Insights”Scaling: Predictive & scheduled: Guides
Section titled “Scaling: Predictive & scheduled: Guides”- Enable predictive scaling
- Forecast and scale a sample workload
- Send metric history to the predictor with OTLP
- Explore and tune forecast models
Scaling: Multi-cluster: Guides
Section titled “Scaling: Multi-cluster: Guides”- Register member clusters with GitOps
- Distribute workload replicas across clusters
- Distribute jobs across clusters
- Scale vCluster workloads centrally
Scaling: Custom targets & node capacity: Guides
Section titled “Scaling: Custom targets & node capacity: Guides”- Prewarm node capacity with Karpenter
- Prewarm node capacity with Cluster Autoscaler
- Prewarm node capacity on GKE
Operate
Section titled “Operate”- Manage cluster connections and fleet health
- Configure Kedify for high availability
- Pause and resume scaling for maintenance
- Collect diagnostics for support
- Schedule scaling to resume at a time
- Schedule scaling to resume after a delay
- Install Autoscaling Checks
- Configure Checks with Prometheus
- Configure Checks with OTel
- Verify Autoscaling Checks
- Remove Autoscaling Checks
- Install kubectl kedify
Scaling: Vertical
Section titled “Scaling: Vertical”- Enable and try PRP
- Enable PRA and opt workloads in
- Verify PRA and inspect resource events
- Pause and resume PRA
Scaling: Horizontal scaling
Section titled “Scaling: Horizontal scaling”- Set up KPA
- Select KPA for a ScaledObject
- Tune KPA evaluation interval
- Roll back KPA to HPA
- Configure a ScalingGroup