Skip to content

For the complete documentation index and AI-optimized content, see /llms.txt. All pages support markdown format via .md extension or Accept: text/markdown header.

Best Practices

For the complete documentation index and AI-optimized content, see /llms.txt. All pages support markdown format via .md extension or Accept: text/markdown header.

Use these guides when you already know which scaler or Kedify feature you want to use, and you want the workload and cluster behavior around it to be stable in production.

Start with the Kubernetes autoscaling guide for workload-level behavior, then tune the KEDA configuration that drives the scaling signal.

Run autoscaling checks in production-like clusters to verify that the autoscaling control plane works end to end. The checks exercise HTTP, Prometheus, OTel, CPU, and memory scaling paths, then expose kedify_autoscaling_check_duration_seconds as the primary alerting metric for the latest check iteration latency.

Alert when the latest iteration takes longer than your operating threshold, and use the step status labels and step duration metrics to identify whether the delay is in signal generation, metric availability, pod-autoscaler activity, scale-up, or scale-down.