Add activity log system for provisioning lifecycle visibility

Hosts stuck in states like pxe_ready had zero visibility into why.
This adds a persistent activity log that records every meaningful
step (state transitions, PXE events, cluster join stages, failures)
and surfaces it on the host detail page with live SSE updates.
Includes a stuck-detection warning banner when a host sits in
pxe_ready for >10 minutes with no iPXE request.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

internal/api/ui.go

@@ -24,6 +24,7 @@ type UI struct {
 	Ops          *store.Operations
 	Locks        *store.Locks
 	Images       *store.Images
+	Activity     *store.Activity
 	ImageSvc     *image.Service
 	Runner       *orchestrator.Runner
 	Orchestrator *orchestrator.HostOrchestrator
@@ -106,7 +107,8 @@ func (u *UI) HostDetail(w http.ResponseWriter, r *http.Request) {
 		return
 	}
 	ops, _ := u.Ops.ListByHost(r.Context(), host.ID)
-	renderHTML(w, hostDetailPage(host, ops))
+	activity, _ := u.Activity.ListByHost(r.Context(), host.ID, 50)
+	renderHTML(w, hostDetailPage(host, ops, activity))
 }
 
 func (u *UI) TriggerRebuild(w http.ResponseWriter, r *http.Request) {
@@ -131,6 +133,7 @@ func (u *UI) TriggerRebuild(w http.ResponseWriter, r *http.Request) {
 		Kind:   model.OpRebuildProxmox,
 	})
 	_ = u.Locks.Acquire(r.Context(), host.ID, opID)
+	u.Runner.LogActivity(r.Context(), host.ID, model.LogInfo, "ui", "Rebuild triggered by user")
 
 	if err := u.Orchestrator.PrepareRebuild(r.Context(), host.ID); err != nil {
 		_ = u.Locks.Release(r.Context(), host.ID)