fix(plex): retry slow library hubs

lib/services/image_cache_service.dart

@@ -6,6 +6,8 @@ import 'package:http/http.dart' as http;
 
 import '../utils/media_server_http_client.dart';
 
+final _artworkHttpClient = MediaServerHttpClient(usePlexApiClient: true);
+
 /// Shared cache manager for media-server image artwork. Used for both Plex and
 /// Jellyfin artwork (the class name predates Jellyfin support — it's
 /// backend-neutral).
@@ -13,7 +15,8 @@ import '../utils/media_server_http_client.dart';
 /// Uses the platform-native HTTP client so iOS/macOS (CupertinoClient) and
 /// Android (CronetClient) benefit from HTTP/2 connection multiplexing —
 /// many concurrent image downloads over a single connection instead of
-/// being limited to a handful of HTTP/1.1 connections.
+/// being limited to a handful of HTTP/1.1 connections. On Linux this uses the
+/// same finite-connection tuning as Plex API traffic.
 class PlexImageCacheManager extends ce_cache.DefaultCacheManager {
   static final PlexImageCacheManager instance = PlexImageCacheManager._();
 
@@ -21,7 +24,7 @@ class PlexImageCacheManager extends ce_cache.DefaultCacheManager {
     : super(
         stalePeriod: const Duration(days: 14),
         maxNrOfCacheObjects: 3000,
-        httpClientFactory: () => _SharedHttpClient(httpClient.inner),
+        httpClientFactory: () => _SharedHttpClient(_artworkHttpClient.inner),
       );
 }
 

lib/services/plex_client.dart

@@ -1542,10 +1542,16 @@ class PlexClient with MediaServerCacheMixin, _PlexLiveTvClientMethods implements
   /// Returns a list of recommendation hubs like "Trending Movies", "Top in Genre", etc.
   Future<List<PlexHubDto>> _getLibraryHubs(String sectionId, {int limit = 10}) async {
     try {
-      final response = await _getWithFailover(
-        '/hubs/sections/$sectionId',
-        queryParameters: {'count': limit, 'includeGuids': 1},
-        allowEndpointFailover: false,
+      final response = await retryTransientMediaServerCall(
+        operation: 'Plex library hubs',
+        attemptTimeouts: MediaServerTimeouts.libraryHubAttemptTimeouts,
+        call: (timeout, abort) => _getWithFailover(
+          '/hubs/sections/$sectionId',
+          queryParameters: {'count': limit, 'includeGuids': 1},
+          timeout: timeout,
+          abort: abort,
+          allowEndpointFailover: false,
+        ),
       );
       final sid = serverId;
       final sname = serverName;

lib/utils/media_server_timeouts.dart

@@ -11,6 +11,10 @@ class MediaServerTimeouts {
   /// while Plex wakes idle disks, but should not block forever.
   static const homeHubAttemptTimeouts = [Duration(seconds: 10), Duration(seconds: 5), Duration(milliseconds: 2500)];
 
+  /// Retry budget for per-library home hub rows (`/hubs/sections/{id}`). These
+  /// can be slower than the top-level home hub call on remote Plex servers.
+  static const libraryHubAttemptTimeouts = [Duration(seconds: 10), Duration(seconds: 8), Duration(seconds: 5)];
+
   /// Timeout for probing a cached/preferred endpoint before falling back to
   /// the full candidate race (used in [PlexServer.findBestWorkingConnection]).
   static const preferredEndpointProbe = Duration(milliseconds: 1500);