Kedify documentation
Kedify extends KEDA with HTTP, OpenTelemetry and predictive scalers, vertical scaling, and resource recommendations.
What are you trying to achieve?
Section titled “What are you trying to achieve?”| Goal | Start here |
|---|---|
| Reduce compute waste or rightsize workloads | Insights recommendations |
| Dynamically scale pod CPU and memory | Vertical scaling |
| Scale queues, workers or application metrics | Horizontal scaling |
| Scale idle workloads to zero | Scale-to-zero & cold starts |
| Prepare for predictable traffic spikes | Predictive or scheduled scaling |
| Scale across multiple clusters | Distributed scaling |
| Scale a large Kubernetes cluster | Multi-tenant KEDA and KPA |
| Prescale nodes dynamically | Node capacity buffers |
| Scale AI inference | HTTP routing · vLLM metrics |
Get started
Section titled “Get started”Run the quickstart or install with Helm. Already using KEDA? Migrate an existing installation and check KEDA, HPA and Kedify for controller ownership.
Documentation
Section titled “Documentation”| Section | Contents |
|---|---|
| Install | Requirements, Helm/GitOps, versions and restricted environments |
| Operate | Dashboard, tenants, health checks, availability and tuning |
| Scaling | Scalers, controllers and integration recipes |
| Insights | Resource recommendations and FinOps |
| Best Practices | Production configuration, reliability and tuning |
| Security & compliance | Access, data boundaries, FIPS and SOC 2 |