Move this branch's unified-resources note off the end of the file

Every lane appends to the end of this contract, so a note there conflicts with each merge to main. Taking main's file with the note placed beside the Docker sections lets main merge cleanly now and later.

docs/release-control/v6/internal/subsystems/unified-resources.md

@@ -5737,6 +5737,23 @@ A failed admission HTTP refresh is not an empty estate either: retain the last
 valid facet on request failure. A successful response still replaces it,
 including all-false admission; first-load failure remains unresolved.
 
+### Docker tables state silence and drop per-row host context
+
+`EvaluateResourceHealth` keeps the first reason as the one that explains the
+verdict, and when a stronger verdict (an alert, an availability failure, an
+offline host) wins while a source is stale, it appends a trailing
+`telemetry_stale` reason with the backend-formatted age, so the summary API's
+top reason is unchanged while surfaces can still tell that a resource stopped
+reporting. That reason, alone under a `stale` verdict or trailing another, is
+the only staleness signal platform tables read, through
+`features/platformPage/resourceStaleness.ts`; tables never derive staleness
+from `lastSeen` themselves. A stale Docker host dims its CPU, memory and
+disk cells and shows the age under Uptime ("No report for 1h" on hover and
+for assistive tech), and its container group header says the same. In
+grouped-by-host mode the containers table drops the Host and Engine
+columns, which repeat one value per group, and the group header names the
+engine once; flat mode keeps both columns.
+
 ### Existing Docker update and expanded identity text remain readable
 
 DockerContainersTable marks its update cell as the phone reflow boundary for both
@@ -5870,20 +5887,15 @@ presentation only: the action audit list, approval counts, review dialog,
 routing and every recorded action outcome are read and rendered exactly as
 before, and no resource or action identity changes.
 
-### Docker tables state silence and drop per-row host context
+### vSphere Overview leads with vCenter health signals
 
-`EvaluateResourceHealth` keeps the first reason as the one that explains the
-verdict, and when a stronger verdict (an alert, an availability failure, an
-offline host) wins while a source is stale, it appends a trailing
-`telemetry_stale` reason with the backend-formatted age, so the summary API's
-top reason is unchanged while surfaces can still tell that a resource stopped
-reporting. That reason, alone under a `stale` verdict or trailing another, is
-the only staleness signal platform tables read, through
-`features/platformPage/resourceStaleness.ts`; tables never derive staleness
-from `lastSeen` themselves. A stale Docker host dims its CPU, memory and
-disk cells and shows the age under Uptime ("No report for 1h" on hover and
-for assistive tech), and its container group header says the same. In
-grouped-by-host mode the containers table drops the Host and Engine
-columns, which repeat one value per group, and the group header names the
-engine once; flat mode keeps both columns.
+The vSphere Overview answers "is anything in trouble?" before inventory, as
+the TrueNAS Overview does. While the Overview is active the page also
+hydrates the Health tab's resource query (hosts, VMs, datastores and
+networks) and, when `buildVmwareIncidentRows` returns any rows from those
+resources' canonical `incidents`, renders the shared `VsphereAlertsTable`
+above the hosts table without its filter toolbar. Nothing renders when
+vCenter reports no signals. The Overview's own model, workload snapshot and
+navigation facets still come from the Overview query; the Health tab keeps
+the filterable table.