fix(deployment): instance delete fully removes the record

Deleting an instance only undeployed it from the site and set the state
to NotDeployed, leaving an orphan record that could never be removed —
the state-transition matrix rejected delete from NotDeployed.

Delete now removes the instance record entirely (deployment history,
snapshot, attribute/alarm overrides, and connection bindings go with
it), and is permitted from any state.

src/ScadaLink.DeploymentManager/DeploymentService.cs

@@ -282,7 +282,9 @@ public class DeploymentService
     }
 
     /// <summary>
-    /// WP-6: Delete an instance. Stops actor, removes config. S&amp;F NOT cleared.
+    /// WP-6: Delete an instance. Stops the site actor, removes site config, and
+    /// removes the central instance record (deployment history, snapshot,
+    /// overrides, and connection bindings go with it). S&amp;F NOT cleared.
     /// Delete fails if site unreachable (30s timeout via CommunicationOptions).
     /// </summary>
     public async Task<Result<InstanceLifecycleResponse>> DeleteInstanceAsync(
@@ -309,12 +311,10 @@ public class DeploymentService
 
         if (response.Success)
         {
-            // Remove deployed snapshot
-            await _repository.DeleteDeployedSnapshotAsync(instanceId, cancellationToken);
-
-            // Set state to NotDeployed (or the instance record could be deleted entirely by higher layers)
-            instance.State = InstanceState.NotDeployed;
-            await _repository.UpdateInstanceAsync(instance, cancellationToken);
+            // Delete means delete: remove the instance record entirely.
+            // Deployment records, snapshot, overrides, and connection bindings
+            // are removed with it (see repository implementation).
+            await _repository.DeleteInstanceAsync(instanceId, cancellationToken);
             await _repository.SaveChangesAsync(cancellationToken);
         }