Recommendations now respect container-level HPA targets

PerfectScale now recognizes HPA triggers that target a single container, keeping right-sizing recommendations accurate for workloads using container-scoped autoscaling. Recommendations apply only to the container the HPA watches, preventing over- or under-sized suggestions. Podfit and Zoom-In now distinguish pod-level from container-level triggers, so the data you see matches how your autoscaler actually behaves.

PerfectScale now detects and displays HPA triggers of type ContainerResource, so right-sizing recommendations stay accurate when your HPA scales a single container instead of the whole Pod. No more over- or under-sized recommendations for workloads using container-scoped autoscaling.

What changed

Smarter recommendations for container-scoped HPA

Kubernetes lets you target HPA at a single container instead of the whole Pod. PerfectScale now reads the HPA target scope, so recommendations only apply to the specific container the HPA watches.

More accurate data collection

Under the hood, we read the container label that Kube State Metrics exposes on HPA target metrics, so trigger data is collected per container \u2014 meaning the numbers you see in the UI match the container the HPA actually watches.

UI updates

Podfit

HPA trigger data now distinguishes between pod-level and container-level metrics.

Zoom-In

The HPA Triggers panel now groups triggers by scope and surfaces related alerts at the top of the panel.