Hosting: read-only InstanceAction ServerEvents to land Azure activity-log and Resource Health events for a deployment's databaseServer #7379

MeshWeaver#7068: memex's managed Postgres flexible server memexaks-pg restarted or failed over at 2026-10-22 21:22:51Z. All sessions were cut at once on every pod, then connections were refused for about 25–40 s (governed Logs Ops/Actions/i7068-pg-terminated-memex-20261023 and i7068-pg-refused-memexcloud-tight-20261023 on control). The CAUSE (planned maintenance, an HA failover or a crash) is only in Azure's activity log and Resource Health for the server. No governed InstanceAction can read them, so the cause cannot be established without break-glass az access, which is not allowed.

Wanted: a read-only InstanceAction (e.g. requestedAction 'ServerEvents') that, for a deployment's declared databaseServer, lands the activity-log entries and Resource Health transitions for a time window as nodes, the same way Logs lands Loki lines. It also needs to report whether high availability is enabled on the server.

Not established: whether the operator's managed identity already holds Reader on the server's resource group.

Context: maintainer-designated follow-up to #7068 (decision 2026-10-23: #7068 downgraded to sev:M, with this follow-up). Where: Deployments/memex — the control instance memex.systemorph.com, whose declared databaseServer is memexaks-pg (flexible server, resource group memex-aks-rg). The gap this feature would close was measured live: at 2026-10-22 21:22:51Z every pod's sessions were cut at once and connections were refused for 25–40 s, and the cause could not be established from any governed surface — only Azure's activity log and Resource Health hold it, and no InstanceAction can read them.

Category: idea (feature request — read-only observability InstanceAction; no bug, no sev).

Origin: https://control.systemorph.com/Feedback/_Submissions/fb-389bb17351c45b9b9b71f74aed11a5aa

filed by the triage agent on behalf of an unattributed submitter via systemorph-com

Triage routing: filed in Systemorph/MeshWeaver — agent: Systemorph/MeshWeaver — Owner resolved to Systemorph/MeshWeaver by evidence: the feedback names MeshWeaver#7068 and asks for a new read-only InstanceAction (requestedAction 'ServerEvents') in the Hosting operator — platform code whose InstanceAction/Reconcile machinery and defects are tracked in MeshWeaver core (e.g. #3781 hosting-deploy, #3201 RegistryTokenResolver). The Space of the feedback's node (Deployments) is GitSynced from Systemorph/Memex, but that repository holds this deployment's mesh records and AKS config, not the InstanceAction code this feature extends..