Kubernetes overview dashboards

This release introduces the first version of the Overview pages — two at-a-glance dashboards for a cluster and for its workloads — and a new Policy section in the sidebar gathering the kinds that constrain what a namespace is allowed to do.

📊 Cluster & Workloads Overviews (new)

Two new dashboard pages, reachable from the sidebar, the command palette, or a deeplink.

• Cluster overview: nodes and capacity, a Requests/Allocatable band, control-plane info, recent warning events, hot nodes, volumes, certificates and custom resources — each summarised as a health card
• Workloads overview: pooled health cards per workload kind (Deployments, StatefulSets, DaemonSets, Jobs, CronJobs…), plus sections for recent restarts, pods under pressure, and pods stuck in init or termination
• Health cards: one problem-first line per card, a health bar with the full status breakdown, and a tooltip naming exactly why a card is unhealthy — "All clear" in green when there's nothing to report
• Drill down: every figure on a card is a link into the matching filtered resource list; cards themselves open in a tab, a panel or a new window
• Shared health taxonomy: a single status classification used across Pods, workloads, Namespaces, PersistentVolumes, Certificates and CronJobs (including a paused state), also exposed as a sortable _health column in resource lists
• Multi-cluster: works under the "all clusters" selection, with a context colour dot on each cluster card

🛡️ Policy Section

A new Policy sidebar group collecting the kinds that constrain a namespace rather than describe what it runs — with four new resource modules.

• PodDisruptionBudgets: dedicated list columns, an overview with a disruption headroom bar, a Pods tab listing the pods it covers, and a Set Budget action
• HorizontalPodAutoscalers: new module with target progress bars per metric, scale target and replica range, and a Set Replica Range action
• NetworkPolicies: dedicated list columns, a policy-flow overview (ingress/egress peers and ports, isolation state), and a Pods tab
• ResourceQuotas and LimitRanges: two new modules, each rolled up to a single usage/state summary
• Drain safety: draining a node now warns you up front about the disruption budgets that would block it, and tells "no blockers" apart from "couldn't check"

✨ Other Improvements

• Port forwarding: open the port-forward manager with the G P shortcut
• Sidebar: larger mega-group section titles for easier scanning

🐛 Fixes

• Flux: HelmRelease and Kustomization status is now derived from the Ready condition first
• Tab bar: the panel picker no longer gets clipped by the glass chrome
• HPA: nano/micro quantity suffixes are parsed correctly, and the replica band's track is visible again
• Overview: pod-signal counts are correct, a failed Events watch is reported instead of silently showing "no warnings", and the custom-resources card loads under the "all clusters" selection
• Routing: the panel selection is preserved when leaving the resources view