fix: ops pages no longer 500 or hang when RabbitMQ management isn't configured

SW.Bitween.Api/Resources/Ops/UnattendedQueues.cs

@@ -4,6 +4,7 @@
 using System.Text.RegularExpressions;
 using System.Threading.Tasks;
 using EasyNetQ.Management.Client;
+using EasyNetQ.Management.Client.Model;
 using Microsoft.Extensions.Caching.Memory;
 using SW.Bitween.Domain;
 using SW.Bus;
@@ -59,13 +60,23 @@ public async Task<object> Handle()
         }
 
         // Cached on the same clock as the bus's own management call, because this page polls.
-        var queues = await memoryCache.GetOrCreateAsync("bitween-all-queues", async entry =>
+        // A failed fetch is deliberately not cached: caching it would report "no queues" for the
+        // rest of the cache window even after the management API recovers.
+        if (!memoryCache.TryGetValue("bitween-all-queues", out IReadOnlyList<Queue> queues))
         {
-            entry.AbsoluteExpirationRelativeToNow = TimeSpan.FromSeconds(busOptions.MonitoringCacheSeconds);
-            var client = new ManagementClient(new Uri(busOptions.ManagementUrl),
-                busOptions.ManagementUsername, busOptions.ManagementPassword);
-            return await client.GetQueuesAsync(busOptions.VirtualHost);
-        });
+            try
+            {
+                var client = new ManagementClient(new Uri(busOptions.ManagementUrl),
+                    busOptions.ManagementUsername, busOptions.ManagementPassword);
+                queues = await client.GetQueuesAsync(busOptions.VirtualHost);
+                memoryCache.Set("bitween-all-queues", queues, TimeSpan.FromSeconds(busOptions.MonitoringCacheSeconds));
+            }
+            catch
+            {
+                // Management API unreachable or misconfigured - degrade to "no data" instead of 500ing.
+                queues = Array.Empty<Queue>();
+            }
+        }
 
         var orphans = queues
             .Where(q => q.Name.StartsWith($"{prefix}.", StringComparison.OrdinalIgnoreCase))

SW.Bitween.Api/Resources/WorkGroups/Search.cs

@@ -43,6 +43,10 @@ public async Task<object> Handle(SearchWorkGroupModel request)
         }
         catch (Exception ex) when (ex is TaskCanceledException or TimeoutException or System.Net.Http.HttpRequestException)
         {
+            // The RabbitMQ management API call itself already degrades to empty data on failure
+            // (ConsumerReader.GetConsumerCounts) - this only guards the network-facing edge of
+            // that call. Anything else (e.g. ConsumerDiscovery.Load() throwing) is a real bug and
+            // should surface as a 500, not silently render as "no metrics".
             logger.LogWarning(ex, "Unable to load RabbitMQ consumer metrics for work groups.");
         }
 

SW.Bitween.Api/SW.Bitween.Api.csproj

@@ -32,8 +32,8 @@
          them to ask the broker what exists; without this it would have to rebuild each of them
          from configuration and get them subtly wrong. Already in the app through SW.Bitween.Web —
          keep the version in step with it. -->
-    <PackageReference Include="SimplyWorks.Bus" Version="8.1.11" />
-    <PackageReference Include="SimplyWorks.Bus.RabbitMqExtensions" Version="8.1.11" />
+    <PackageReference Include="SimplyWorks.Bus" Version="8.1.18" />
+    <PackageReference Include="SimplyWorks.Bus.RabbitMqExtensions" Version="8.1.18" />
     <PackageReference Include="SimplyWorks.EfCoreExtensions" Version="8.1.2" />
     <PackageReference Include="SimplyWorks.Scheduler.Sdk" Version="8.1.7" />
   </ItemGroup>

SW.Bitween.IntegrationTests/SW.Bitween.IntegrationTests.csproj

@@ -15,7 +15,7 @@
 
   <ItemGroup>
     <PackageReference Include="Microsoft.NET.Test.Sdk" Version="17.12.0" />
-    <PackageReference Include="SimplyWorks.CloudFiles.LocalTests.Extensions" Version="8.1.6" />
+    <PackageReference Include="SimplyWorks.CloudFiles.LocalTests.Extensions" Version="8.1.12" />
     <PackageReference Include="xunit" Version="2.9.2" />
     <PackageReference Include="xunit.runner.visualstudio" Version="2.8.2">
       <PrivateAssets>all</PrivateAssets>
@@ -27,10 +27,10 @@
     <PackageReference Include="Testcontainers.RabbitMq" Version="3.10.0" />
 
     <!-- Serverless runtime -->
-    <PackageReference Include="SimplyWorks.Serverless" Version="8.1.5" />
+    <PackageReference Include="SimplyWorks.Serverless" Version="8.1.11" />
 
     <!-- Bus (for IPublish + AddBus registration) -->
-    <PackageReference Include="SimplyWorks.Bus" Version="8.1.11" />
+    <PackageReference Include="SimplyWorks.Bus" Version="8.1.18" />
 
     <!-- EF Core naming conventions (used by PgSql context) -->
     <PackageReference Include="EFCore.NamingConventions" Version="9.0.0" />

SW.Bitween.NativeAdapters/S3Receiver/NativeS3Receiver.cs

@@ -13,7 +13,7 @@ public class NativeS3Receiver : INativeInfolinkReceiver, IDisposable
 
     public Task Initialize()
     {
-        var options = new CloudFilesOptions
+        var options = new S3CloudFilesOptions
         {
             AccessKeyId = _options.AccessKeyId,
             SecretAccessKey = _options.SecretAccessKey,

SW.Bitween.NativeAdapters/SW.Bitween.NativeAdapters.csproj

@@ -23,7 +23,7 @@
       <PackageReference Include="Rebex.Pop3" Version="8.0.9673" />
       <PackageReference Include="Rebex.Sftp" Version="8.0.9673" />
       <PackageReference Include="Scriban" Version="7.0.6" />
-      <PackageReference Include="SimplyWorks.CloudFiles.S3" Version="8.1.1" />
+      <PackageReference Include="SimplyWorks.CloudFiles.S3" Version="8.1.12" />
       <PackageReference Include="SimplyWorks.PrimitiveTypes" Version="8.1.5" />
     </ItemGroup>
 

SW.Bitween.SampleConfigurableAdapter/SW.Bitween.SampleConfigurableAdapter.csproj

@@ -5,6 +5,6 @@
     <RootNamespace>SW.Bitween.SampleConfigurableAdapter</RootNamespace>
   </PropertyGroup>
   <ItemGroup>
-    <PackageReference Include="SimplyWorks.Serverless.Sdk" Version="8.1.1" />
+    <PackageReference Include="SimplyWorks.Serverless.Sdk" Version="8.1.11" />
   </ItemGroup>
 </Project>

SW.Bitween.SampleHandler/SW.Bitween.SampleHandler.csproj

@@ -7,7 +7,7 @@
   </PropertyGroup>
 
   <ItemGroup>
-    <PackageReference Include="SimplyWorks.Serverless.Sdk" Version="8.1.1" />
+    <PackageReference Include="SimplyWorks.Serverless.Sdk" Version="8.1.11" />
   </ItemGroup>
 
 

SW.Bitween.SampleMapper/SW.Bitween.SampleMapper.csproj

@@ -7,7 +7,7 @@
   </PropertyGroup>
 
   <ItemGroup>
-    <PackageReference Include="SimplyWorks.Serverless.Sdk" Version="8.1.1" />
+    <PackageReference Include="SimplyWorks.Serverless.Sdk" Version="8.1.11" />
   </ItemGroup>
 
 </Project>

SW.Bitween.SampleValidator/SW.Bitween.SampleValidator.csproj

@@ -8,7 +8,7 @@
 
   <ItemGroup>
     <PackageReference Include="FluentValidation" Version="11.11.0" />
-    <PackageReference Include="SimplyWorks.Serverless.Sdk" Version="8.1.1" />
+    <PackageReference Include="SimplyWorks.Serverless.Sdk" Version="8.1.11" />
   </ItemGroup>
 
 

SW.Bitween.UnitTests/SW.Bitween.UnitTests.csproj

@@ -19,7 +19,7 @@
     <PackageReference Include="MSTest.TestAdapter" Version="3.7.1" />
     <PackageReference Include="MSTest.TestFramework" Version="3.7.1" />
     <PackageReference Include="SimplyWorks.CqApi" Version="8.2.6" />
-    <PackageReference Include="SimplyWorks.Bus" Version="8.1.11" />
+    <PackageReference Include="SimplyWorks.Bus" Version="8.1.18" />
   </ItemGroup>
 
   <ItemGroup>

SW.Bitween.Web/ClientApp/src/api/http/dashboard.ts

@@ -45,7 +45,9 @@ export const dashboardMethods = {
         `/xchanges?filter=${encodeURIComponent(`StartedOn:6:${new Date(windowStart).toISOString()}`)}&size=1000&sort=StartedOn:1`,
       ),
       get<SearchyResponse<unknown>>("/delayedretries?size=1"),
-      get<RawAlert[]>("/ops/alerts"),
+      // Depends on RabbitMQ management being configured on the backend — don't let it take the
+      // rest of the dashboard down when it isn't; the "Queue alerts" tile flags it instead.
+      get<RawAlert[]>("/ops/alerts").catch(() => null),
       subscriptionMethods.listSubscriptionRows(),
     ]);
 
@@ -117,7 +119,7 @@ export const dashboardMethods = {
       yesterdayTotal: yesterdayRows.length,
       successRate7d,
       pendingRetries: delayedRes.totalCount,
-      queueAlerts: alertsRaw.filter((a) => a.severity !== "Info").length,
+      queueAlerts: alertsRaw === null ? null : alertsRaw.filter((a) => a.severity !== "Info").length,
       trafficByDay,
       busiest,
       latestFailures,

SW.Bitween.Web/ClientApp/src/api/types.ts

@@ -1002,7 +1002,8 @@ export interface DashboardData {
   /** Percentage 0–100 across the last 7 days of finished exchanges. */
   successRate7d: number;
   pendingRetries: number;
-  queueAlerts: number;
+  /** Null when the RabbitMQ management API is unavailable — never counts as zero. */
+  queueAlerts: number | null;
   /** Last 14 days, oldest first; today is the final entry. */
   trafficByDay: { date: string; success: number; failed: number }[];
   /** Top subscriptions by 7-day traffic, busiest first. */

SW.Bitween.Web/ClientApp/src/components/ui/basics.tsx

@@ -1,5 +1,5 @@
 import type { ButtonHTMLAttributes, ReactNode } from "react";
-import { Loader2 } from "lucide-react";
+import { Loader2, TriangleAlert } from "lucide-react";
 
 type ButtonVariant = "primary" | "secondary" | "ghost" | "danger";
 
@@ -108,6 +108,16 @@ export function EmptyState({
   );
 }
 
+/** Slim inline banner for a known, expected degraded state (e.g. an optional integration not configured). */
+export function InlineNotice({ children }: { children: ReactNode }) {
+  return (
+    <p className="mb-3 flex items-start gap-1.5 rounded-lg border border-warn-100 bg-warn-100/40 px-3 py-2 text-[13px] text-warn-800">
+      <TriangleAlert className="mt-0.5 size-3.5 shrink-0" aria-hidden />
+      <span>{children}</span>
+    </p>
+  );
+}
+
 /** Inline error strip for failed form submissions. */
 export function FormError({ children }: { children: ReactNode }) {
   if (!children) return null;

SW.Bitween.Web/ClientApp/src/lib/appConfig.ts

@@ -0,0 +1,13 @@
+import { useQuery } from "@tanstack/react-query";
+import { getAppConfig } from "../api";
+
+/** Same query key as `useBranding`, so this shares its cache rather than firing a second fetch. */
+export function useAppConfig() {
+  return useQuery({ queryKey: ["appConfig"], queryFn: getAppConfig });
+}
+
+/** Defaults to true while the config is still loading, so pages don't flash a "not configured" state. */
+export function useRabbitMqManagementConfigured(): boolean {
+  const { data } = useAppConfig();
+  return data?.isRabbitMqManagementConfigured ?? true;
+}

SW.Bitween.Web/ClientApp/src/pages/dashboard/DashboardPage.tsx

@@ -5,6 +5,7 @@ import { api } from "../../api";
 import { PageHeader } from "../../components/layout/PageHeader";
 import { Badge, EmptyState, LoadingBlock } from "../../components/ui/basics";
 import { Panel } from "../../components/ui/Panel";
+import { useRabbitMqManagementConfigured } from "../../lib/appConfig";
 import { timeAgo } from "../../lib/dates";
 import { StatusBadge, XchangeId } from "../exchanges/shared";
 
@@ -50,15 +51,29 @@ function StatTile({
  * what needs a human — everything links into the page that can act on it.
  */
 export function DashboardPage() {
-  const { data, isLoading } = useQuery({
+  const rabbitMqConfigured = useRabbitMqManagementConfigured();
+  const { data, isLoading, isError } = useQuery({
     queryKey: ["dashboard"],
     queryFn: () => api.getDashboard(),
     refetchInterval: 60_000,
     placeholderData: keepPreviousData,
   });
 
+  if (isError && !data)
+    return (
+      <EmptyState title="Couldn't load the dashboard">
+        Something went wrong fetching dashboard data. It'll keep retrying in the background — try
+        refreshing the page.
+      </EmptyState>
+    );
   if (isLoading || !data) return <LoadingBlock label="Putting the picture together…" />;
 
+  const queueAlerts = rabbitMqConfigured ? data.queueAlerts : null;
+  const queueAlertsSub = !rabbitMqConfigured
+    ? "needs RabbitMQ management configured"
+    : queueAlerts === null
+      ? "unavailable right now"
+      : "live consumer health";
   const delta = data.today.total - data.yesterdayTotal;
   const maxDay = Math.max(1, ...data.trafficByDay.map((d) => d.success + d.failed));
   const needsAttention =
@@ -107,10 +122,10 @@ export function DashboardPage() {
         />
         <StatTile
           label="Queue alerts"
-          value={data.queueAlerts}
-          sub="live consumer health"
+          value={queueAlerts ?? "—"}
+          sub={queueAlertsSub}
           to="/queue-health"
-          accent={data.queueAlerts > 0 ? "warn" : undefined}
+          accent={queueAlerts !== null && queueAlerts > 0 ? "warn" : undefined}
         />
       </div>
 

SW.Bitween.Web/ClientApp/src/pages/queue-health/QueueHealthPage.tsx

@@ -1,13 +1,14 @@
 import type { ReactNode } from "react";
 import { Link } from "react-router";
 import { keepPreviousData, useQuery } from "@tanstack/react-query";
-import { AlertTriangle, OctagonAlert } from "lucide-react";
+import { AlertTriangle, OctagonAlert, Unplug } from "lucide-react";
 import { api, type ConsumerHealth, type QueueLane, type QueueSeverity } from "../../api";
 import { PageHeader } from "../../components/layout/PageHeader";
 import { Badge, EmptyState, LoadingBlock } from "../../components/ui/basics";
 import { queueHealthTitle } from "../../components/config/shared";
 import { Panel } from "../../components/ui/Panel";
 import { timeAgo } from "../../lib/dates";
+import { useRabbitMqManagementConfigured } from "../../lib/appConfig";
 
 const POLL_MS = 5_000;
 
@@ -89,13 +90,38 @@ function StatTile({ label, value, sub }: { label: string; value: ReactNode; sub?
  * stays on screen while the next one loads, so nothing flashes.
  */
 export function QueueHealthPage() {
-  const { data, isLoading, dataUpdatedAt } = useQuery({
+  const rabbitMqConfigured = useRabbitMqManagementConfigured();
+  const { data, isLoading, isError, dataUpdatedAt } = useQuery({
     queryKey: ["queue-health"],
     queryFn: () => api.getQueueHealth(),
     refetchInterval: POLL_MS,
     placeholderData: keepPreviousData,
+    enabled: rabbitMqConfigured,
   });
 
+  if (!rabbitMqConfigured)
+    return (
+      <div>
+        <PageHeader
+          title="Queue health"
+          description="Live throughput and backlog for every queue this instance consumes."
+        />
+        <EmptyState icon={<Unplug />} title="RabbitMQ management isn't configured">
+          This page reads live queue stats from the RabbitMQ management API, which the backend
+          isn't configured to reach. Ask an admin to set the management URL, username and
+          password, then reload.
+        </EmptyState>
+      </div>
+    );
+
+  if (isError && !data)
+    return (
+      <EmptyState title="Couldn't load queue statistics">
+        Something went wrong reaching the RabbitMQ management API. It'll keep retrying in the
+        background — try refreshing the page.
+      </EmptyState>
+    );
+
   if (isLoading || !data) return <LoadingBlock label="Reading queue statistics…" />;
 
   const { summary, consumers, retryBacklog, deadLetters, unattended, alerts } = data;

SW.Bitween.Web/ClientApp/src/pages/work-groups/LiveQueueStats.tsx

@@ -3,6 +3,7 @@ import { keepPreviousData, useQuery } from "@tanstack/react-query";
 import { api } from "../../api";
 import { Badge } from "../../components/ui/basics";
 import { queueHealthTitle } from "../../components/config/shared";
+import { useRabbitMqManagementConfigured } from "../../lib/appConfig";
 
 function LiveStat({ label, value, tone }: { label: string; value: ReactNode; tone?: "warn" | "danger" }) {
   return (
@@ -25,12 +26,25 @@ function LiveStat({ label, value, tone }: { label: string; value: ReactNode; ton
  * these as columns instead, off the same shared query.
  */
 export function LiveQueueStats({ groupId }: { groupId: number }) {
-  const { data } = useQuery({
+  const rabbitMqConfigured = useRabbitMqManagementConfigured();
+  const { data, isError } = useQuery({
     queryKey: ["queue-health"],
     queryFn: () => api.getQueueHealth(),
     refetchInterval: 5_000,
     placeholderData: keepPreviousData,
+    enabled: rabbitMqConfigured,
   });
+
+  if (!rabbitMqConfigured)
+    return (
+      <p className="text-sm text-ink-500">
+        Live queue stats need RabbitMQ management configured on the backend.
+      </p>
+    );
+
+  if (isError && !data)
+    return <p className="text-sm text-ink-500">Couldn't reach RabbitMQ management for live queue stats.</p>;
+
   const consumer = data?.consumers.find((c) => c.workGroupId === groupId);
 
   if (!consumer)