Kubernetes Inventory Controls and Agent Task Filters
added
This week extends the Kubernetes management release with clearer inventory navigation, richer node and workload detail, and easier Agent task filtering. It also includes backend reliability work for deployment networking and policy resolution across target-scoped Agent tasks.
Kubernetes Inventory Review Improvements
Kubernetes inventory views now make it easier to understand what changed in a cluster and where attention is needed:
- Inventory diffs now highlight replica readiness changes, so teams can quickly see when desired or ready replica counts changed between snapshots.
- Diff highlights animate in the inventory diagram, making added, removed, and changed resources easier to spot during review.
- Cluster nodes now render as a dedicated top band in the inventory diagram, giving the cluster itself a clearer place in the resource map.
- Cluster summary information now calls out exposure and warning signals more directly.
- Inventory headers and snapshot labels better communicate the scope of the snapshot being inspected.
These updates build on Kubernetes inventory snapshots, where teams can compare cluster state over time and inspect the impact of manual or Agent-driven changes.
Better Node, Pod, and Workload Detail
Kubernetes detail panels now surface more operational context without forcing teams to jump between views:
- Node details include warning strips, clearer badging, and a divided warnings list for faster scanning.
- Pod summaries show ready and not-ready counts, plus pod-card readiness state.
- Workloads can now be reviewed by app, with per-app dropdowns and clearer workload counts across namespaces.
- Clicking the Nodes summary item centers the node group in the inventory diagram.
- Node details now include a collapsible raw manifest code viewer for teams who need to inspect the underlying Kubernetes resource definition.
The inventory controls also wrap more cleanly on constrained screens, and node card text no longer clips during hover states.
Agent Task Filtering and Policy Reliability
Agent task views now support task type filters. This makes it easier to narrow Agent work to specific task categories such as Kubernetes cleanup, Optimizer, or Log Analysis.
Target-scoped Agent task policies also resolve environment type more reliably. That keeps policy behavior aligned with the selected project environment when a task is scoped to a specific deployment target.
Deploy Networking Reliability
Hyphen moves more deployment networking orchestration onto its newer provisioning layer, including SSL, backend pools, and load balancer creation. This continues the infrastructure reliability work behind Deploy so cloud networking resources can be created, adopted, and reconciled more consistently across environments.
This release also includes an AWS artifact registry secret fix and improved Agent tool selection on the backend, reducing ambiguity when Agent decides which tool should handle a request.