Agentic autoscaler predicts traffic to scale Kubernetes deployments before load arrives
This tool is a Kubernetes operator that polls Prometheus for recent request rates, feeds the data to a forecast service, and patches the target Deployment’s /scale subresource ahead of time. It’s aimed at operators and developers who need to keep latency low during traffic spikes, offering a measurable advantage over the default HPA by reacting before the load hits.
Visit author’s GitHub →Pratyush-Ghosh27/agentic-autoscaler