fix(sentinel): reduce resource churn from health flaps

Ignore health status changes in Sentinel push deduplication when the container lifecycle state is unchanged.

Scope stale resource checks to Sentinel servers whose heartbeat is stale, and avoid refreshing resource last_online_at on unchanged statuses.

app/Http/Controllers/Api/SentinelController.php

@@ -143,11 +143,13 @@ class SentinelController extends Controller
     /**
      * Build a stable hash of container state.
      *
-     * Covers [name, state, health_status] only — metrics and
-     * filesystem_usage_root are excluded on purpose (disk % churns constantly
-     * and would defeat the hash; the storage check is separately cache-gated
-     * inside PushServerUpdateJob). Sorted by name so container ordering from
-     * Sentinel does not affect the hash.
+     * Covers [name, state] only — metrics, filesystem_usage_root, and
+     * health_status are excluded on purpose. Disk % churns constantly, and
+     * health checks can flap between starting/healthy/unhealthy while the
+     * container lifecycle state remains unchanged. Both would otherwise defeat
+     * the hash and dispatch DB-heavy PushServerUpdateJob instances too often.
+     * The force window still refreshes full state periodically. Sorted by name
+     * so container ordering from Sentinel does not affect the hash.
      */
     private function containerStateHash(array $data): string
     {
@@ -155,7 +157,6 @@ class SentinelController extends Controller
             ->map(fn ($c) => [
                 'name' => data_get($c, 'name'),
                 'state' => data_get($c, 'state'),
-                'health_status' => data_get($c, 'health_status'),
             ])
             ->sortBy('name')
             ->values()