Release 0.5¶
16 July 2026 · Full changelog
Release 0.5 adds two complementary views of cluster activity: Kubernetes resource usage from metrics-server and pod-to-pod traffic from Microsoft Retina. Together they make it possible to move from “what is unhealthy?” to “what is consuming resources or talking on the network?” without leaving Kubus.
Cluster and workload metrics¶
The Metrics page charts CPU and memory for the cluster and highlights the busiest nodes, namespaces and pods. Resource lists and detail views use the same samples to show live usage beside requests and limits, so pressure is visible in the context where you are already working.
Metrics come from the Kubernetes Metrics API. Kubus detects whether metrics-server is available and keeps the rest of the application usable when it is not.
Network visibility with Retina¶
The Network Metrics page shows traffic between workloads, namespaces, Services and external addresses. Charts summarize transmitted bytes and packets, while the connection table lets you filter down to individual flows.
Kubus can install a pinned Microsoft Retina deployment into kube-system, configure the
namespaces it observes and remove that installation later. Existing Retina installations
are detected and left under their current owner's control.
Interface polish¶
Tab controls remain visible across light and dark themes, and the documentation screenshots were refreshed for the new metrics views.