feat: async sandbox lifecycle with Redis Stream events

Replace synchronous RPC-based CP-host communication for sandbox
lifecycle operations (Create, Pause, Resume, Destroy) with an async
pattern. CP handlers now return 202 Accepted immediately, fire agent
RPCs in background goroutines, and publish state events to a Redis
Stream. A background consumer processes events as a fallback writer.

Agent-side auto-pause events are pushed to the CP via HTTP callback
(POST /v1/hosts/sandbox-events), keeping Redis internal to the CP.

All DB status transitions use conditional updates
(UpdateSandboxStatusIf, UpdateSandboxRunningIf) to prevent race
conditions between concurrent operations and background goroutines.

The HostMonitor reconciler is kept at 60s as a safety net, extended
to handle transient statuses (starting, pausing, resuming, stopping).

Frontend updated to handle 202 responses with empty bodies and render
transient statuses with blue indicators.

db/queries/sandboxes.sql

@@ -72,7 +72,7 @@ ORDER BY created_at DESC;
 UPDATE sandboxes
 SET status       = 'missing',
     last_updated = NOW()
-WHERE host_id = $1 AND status IN ('running', 'starting', 'pending');
+WHERE host_id = $1 AND status IN ('running', 'starting', 'pending', 'pausing', 'resuming', 'stopping');
 
 -- name: UpdateSandboxMetadata :exec
 UPDATE sandboxes
@@ -80,6 +80,30 @@ SET metadata = $2,
     last_updated = NOW()
 WHERE id = $1;
 
+-- name: UpdateSandboxRunningIf :one
+-- Conditionally transition a sandbox to running only if the current status
+-- matches the expected value. Prevents races where a user destroys a sandbox
+-- while the create/resume goroutine is still in-flight.
+UPDATE sandboxes
+SET status = 'running',
+    host_ip = $3,
+    guest_ip = $4,
+    started_at = $5,
+    last_active_at = $5,
+    last_updated = NOW()
+WHERE id = $1 AND status = $2
+RETURNING *;
+
+-- name: UpdateSandboxStatusIf :one
+-- Atomically update status only when the current status matches the expected value.
+-- Prevents background goroutines from overwriting a status that has since changed
+-- (e.g. user destroyed a sandbox while Create was in-flight).
+UPDATE sandboxes
+SET status       = $3,
+    last_updated = NOW()
+WHERE id = $1 AND status = $2
+RETURNING *;
+
 -- name: BulkRestoreRunning :exec
 -- Called by the reconciler when a host comes back online and its sandboxes are
 -- confirmed alive. Restores only sandboxes that are in 'missing' state.