feat(deployment-manager): resolve DeploymentManager-006 — query site deployment state before redeploy and reconcile

Adds DeploymentStateQuery request/response contracts (Commons), a site-side
handler (SiteRuntime), a CommunicationService query method (Communication), and
reconciliation in DeploymentService: when a prior record is InProgress or
Failed-on-timeout, query the site; if it already holds the target revision hash
mark the record Success without re-sending; on query failure fall through to a
normal deploy (site-side stale-rejection is the safety net).

src/ScadaLink.Communication/CommunicationService.cs

@@ -73,6 +73,26 @@ public class CommunicationService
             envelope, _options.DeploymentTimeout, cancellationToken);
     }
 
+    /// <summary>
+    /// DeploymentManager-006: queries a site for the currently-applied deployment
+    /// identity of a single instance. Used by the Deployment Manager before a
+    /// re-deploy to reconcile against the site's actual state. Sent over the
+    /// existing ClusterClient command/control transport; the Ask times out (no
+    /// central buffering) if the site is unreachable, and the caller falls
+    /// through to a normal deploy.
+    /// </summary>
+    public async Task<DeploymentStateQueryResponse> QueryDeploymentStateAsync(
+        string siteId, DeploymentStateQueryRequest request, CancellationToken cancellationToken = default)
+    {
+        _logger.LogDebug(
+            "Sending DeploymentStateQueryRequest to site {SiteId}, instance={Instance}, correlationId={CorrelationId}",
+            siteId, request.InstanceUniqueName, request.CorrelationId);
+
+        var envelope = new SiteEnvelope(siteId, request);
+        return await GetActor().Ask<DeploymentStateQueryResponse>(
+            envelope, _options.QueryTimeout, cancellationToken);
+    }
+
     // ── Pattern 2: Lifecycle ──
 
     public async Task<InstanceLifecycleResponse> DisableInstanceAsync(