Merge branch 'worktree-agent-a27e5c199270a6d63' into arch-review-remediation

src/ZB.MOM.WW.ScadaBridge.Communication/Actors/ISiteCommandTransport.cs

@@ -24,8 +24,10 @@ public interface ISiteCommandTransport
     /// Routes <paramref name="envelope"/>'s message to its site. Any reply the site produces is
     /// delivered to <paramref name="replyTo"/> — for an <c>Ask</c> that is the temporary ask actor
     /// (completing the caller's task); for a <c>Tell</c>-with-sender (the debug bridge) that is the
-    /// originating actor. A message with no route (an unknown site) is warned and dropped so the
-    /// caller's <c>Ask</c> times out — central never buffers for an unreachable site.
+    /// originating actor. A message with no route (an unknown/unconfigured site) fails fast: a
+    /// <see cref="Status.Failure"/> is delivered to <paramref name="replyTo"/> immediately instead of
+    /// letting the caller's <c>Ask</c> burn its full timeout — central still never buffers for an
+    /// unreachable site.
     /// </summary>
     /// <param name="envelope">The site-addressed command envelope.</param>
     /// <param name="replyTo">Where a reply (or a <see cref="Status.Failure"/>) is delivered.</param>

src/ZB.MOM.WW.ScadaBridge.Communication/Grpc/GrpcSiteTransport.cs

@@ -22,9 +22,10 @@ namespace ZB.MOM.WW.ScadaBridge.Communication.Grpc;
 /// <remarks>
 /// <para>
 /// <b>Per-call deadlines match today's Ask timeouts exactly</b> — see <see cref="ResolveDeadline"/>.
-/// Behaviour is otherwise unchanged from the Akka path: an unknown/unconfigured site is warned and
-/// dropped (the caller's Ask times out), and a transport fault surfaces to the caller as a
-/// <see cref="Status.Failure"/>, which the S&amp;F/audit layers already treat as transient.
+/// An unknown/unconfigured site now fails fast: it surfaces to the caller as a
+/// <see cref="Status.Failure"/> immediately (WP1.6) instead of being warned-and-dropped to let the
+/// caller's Ask burn its full timeout. A transport fault takes the same <see cref="Status.Failure"/>
+/// path, which the S&amp;F/audit layers already treat as transient.
 /// </para>
 /// <para>
 /// <b>Cross-node retry is the channel provider's job</b> and happens only on
@@ -79,13 +80,21 @@ public sealed class GrpcSiteTransport : ISiteCommandTransport
                 replyTo.Tell(reply, ActorRefs.NoSender);
             }
         }
-        catch (SiteChannelUnavailableException)
+        catch (SiteChannelUnavailableException ex)
         {
-            // Parity with the removed Akka "no ClusterClient for site" path: warn and drop, so the caller's
-            // Ask times out. Central never buffers.
+            // WP1.6: a known-dead send (no configured channel for this site) used to warn-and-drop,
+            // leaving the caller's Ask to burn its full timeout for a failure we already know about.
+            // Fail fast instead — same Status.Failure completion path a transport fault takes below —
+            // so S&F/audit treat it as transient immediately rather than tens of seconds later.
+            // Central still never buffers for an unreachable site; only the completion timing changed.
             _logger.LogWarning(
-                "No gRPC channel for site {SiteId}; dropping {Message} (caller's Ask will time out)",
+                "No gRPC channel for site {SiteId}; failing {Message} immediately instead of dropping it",
                 envelope.SiteId, envelope.Message.GetType().Name);
+
+            if (!fireAndForget && !replyTo.IsNobody())
+            {
+                replyTo.Tell(new Status.Failure(ex), ActorRefs.NoSender);
+            }
         }
         catch (Exception ex)
         {

src/ZB.MOM.WW.ScadaBridge.Communication/Grpc/SitePairChannelProvider.cs

@@ -8,9 +8,11 @@ namespace ZB.MOM.WW.ScadaBridge.Communication.Grpc;
 
 /// <summary>
 /// Raised when a site has no usable gRPC channel — an unknown site, or a site with neither
-/// <c>GrpcNodeAAddress</c> nor <c>GrpcNodeBAddress</c> configured. The gRPC transport treats this
-/// the way the removed Akka path treated "no ClusterClient for site": warn and drop, so the caller's Ask
-/// times out (central never buffers).
+/// <c>GrpcNodeAAddress</c> nor <c>GrpcNodeBAddress</c> configured. <see cref="GrpcSiteTransport"/>
+/// treats this as a known-dead send: it fails the caller's <c>Ask</c> fast with an
+/// <c>Akka.Actor.Status.Failure</c> wrapping this exception (WP1.6) instead of warning and dropping
+/// the message, so the caller does not burn its full Ask timeout on a failure already known at send
+/// time (central still never buffers).
 /// </summary>
 public sealed class SiteChannelUnavailableException(string siteId)
     : Exception($"No gRPC channel is configured for site '{siteId}'.")