By the infrastructure you scale (i.e. size of a cluster). The agent collects safe cluster-level telemetry (e.g., nodes/CPU) for accurate, usage-based billing.
Yes. Kedify scales on real signals (HTTP/OTel/queues). If GPU utilization is exposed via metrics, Kedify uses those signals to scale.
Yes-Kedify supports dynamic cluster scaling with predictive policies; we`ll scope prerequisites and validate during the POC where feasible.
Enterprise-ready: FIPS builds, SSO via Auth0, and air-gapped installation paths.
Purchase via AWS, GCP, or Red Hat Marketplaces to speed approvals and count toward commits.