From 55cd60ff8511473aa895fcecc1b595f4fe857314 Mon Sep 17 00:00:00 2001 From: rcourtman <8825017+rcourtman@users.noreply.github.com> Date: Mon, 5 Oct 2026 03:49:14 +0100 Subject: [PATCH] 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. --- .../internal/subsystems/unified-resources.md | 42 ++++++++++++------- 1 file changed, 27 insertions(+), 15 deletions(-) diff --git a/docs/release-control/v6/internal/subsystems/unified-resources.md b/docs/release-control/v6/internal/subsystems/unified-resources.md index c254c02fe8..e3dd015b38 100644 --- a/docs/release-control/v6/internal/subsystems/unified-resources.md +++ b/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.