Move host-agent identity, metric projection, disk/SMART/RAID/Unraid health handling, cleanup, and offline lifecycle into internal/alerts/host.go.
Keep shared health-assessment evaluation package-level for now because storage ZFS and host SMART/RAID still share that bridge, while recording host.go as the host checker owner in the alerts subsystem contract.
Proof: go test ./internal/alerts/...
Move Proxmox node metric, temperature, offline lifecycle, host-agent deduplication, and node display-name cache support into internal/alerts/node.go.
Keep the Manager API unchanged while recording the node checker owner in the alerts subsystem contract and adding a focused display-name cache key characterization.
Proof: go test ./internal/alerts/...
Move PBS connectivity and metric evaluation into internal/alerts/pbs.go, and move storage connectivity, usage, and ZFS health evaluation into internal/alerts/storage.go.
Keep the Manager API unchanged while recording PBS and storage as resource-checker owners in the alerts subsystem contract, with focused characterization tests for PBS offline normalization and ZFS device labels.
Proof: go test ./internal/alerts/...
Move Docker host connectivity, container state and health, metric projection, service gap/update-state checks, image update timing, and Docker tracking cleanup into internal/alerts/docker.go.
Keep the Manager API unchanged while recording Docker as the resource-checker owner and strengthening Docker resource ID normalization proof.
Proof: go test ./internal/alerts/...
Move PMG connectivity, queue, per-node queue, quarantine, and anomaly evaluation into internal/alerts/pmg.go while keeping the Manager API unchanged.
Record PMG as the resource-checker owner in the alerts contract and add a PMG connection-health normalization proof through CheckPMG.
Proof: go test ./internal/alerts/...
Proof: go test ./internal/monitoring
Move active-alert projection, sorting, metadata coercion, recently resolved reads, history wrappers, and notify-existing redispatch into internal/alerts/read_model.go.
Keep the Manager API unchanged while recording the read-side alerts contract owner and strengthening resolved-alert clone proof.
Proof: go test ./internal/alerts/...
Move alert dispatch, flapping suppression, quiet-hours suppression, monitor-only suppression, cooldown, and rate-limit policy into internal/alerts/notification_policy.go.
Keep Manager behavior and public API unchanged while recording the new alerts contract owner and adding monitor-only dispatch proof.
Proof: go test ./internal/alerts/...; go test -json ./internal/api -count=1; go test ./internal/api ./internal/monitoring ./internal/ai/... ./internal/websocket (internal/api did not reproduce on isolated rerun; other packages passed in the broad run).
Extract alert config types, normalization, and identity helpers into internal/alerts/config while preserving the existing alerts package API through aliases and wrappers.
Move Manager callback lifecycle state into a same-package callbackBus, keeping public Set/Subscribe methods unchanged.
Harden metrics SQLite artifacts to owner-only permissions and cover permissive umask behavior.
Proof: go test -json ./internal/api -count=1; go test ./internal/alerts/... ./internal/monitoring ./internal/ai/... ./internal/websocket ./internal/config ./pkg/metrics; go test ./internal/alerts/... ./pkg/metrics
The four-step coachmark over the top tabs was a tour pretending to be
guidance: each step duplicated the tab title in one sentence, and the
Reopen control on /settings/system-general spawned a centered panel with
no spotlight target because the tabs only exist on dashboard routes.
Delete the modal, the localStorage dismissal key, the reopen event, the
Reopen row in General settings, and the matching guardrails so the
shared-primitives tests stop pinning the deleted owner split. Drop the
WhatsNew dismissal helpers and addInitScript bypasses from the
integration suite, and the dedicated tour test in
19-telemetry-disclosure.
Keep the buffered EOF compatibility path for OpenAI-compatible streams that omit [DONE] but provide a terminal finish_reason, while rejecting truncated tool-call streams before they can produce executable tool calls.
Refs #1411
Refs #1412
Retire runtime/API/UI monitored-system volume enforcement now that infrastructure monitoring is no longer capped.
Keep only legacy metadata scrubbing and purchase-start compatibility for old max_monitored_systems references.
Rename the remaining preview surface to monitored-system impact and make previews explanatory rather than save-blocking.
Update subsystem contracts and RA7 evidence for the caps-retired invariant.
rcourtman