Ship production optimization M1-M12.5A: queues, metrics, HTTP observability, and vector search reliability.

Keep similar-ai from tripping the global circuit on a lone URL 502, clamp Qdrant search to 100, and add Server-Timing plus slow-request logging. Studio shared props, Academy S3 exists caching, heat chunking, and Redis/scheduler hygiene stay in this rollout.

config/collections.php

@@ -83,6 +83,9 @@ return [
             'low_engagement_after_days' => (int) env('COLLECTIONS_V5_LOW_ENGAGEMENT_AFTER_DAYS', 14),
         ],
         'queue' => [
+            // No Horizon supervisor consumes this queue today. Keep dispatch
+            // off until a consumer exists (M7). Existing jobs are not deleted.
+            'dispatch_enabled' => (bool) env('COLLECTIONS_V5_DISPATCH_ENABLED', false),
             'name' => env('COLLECTIONS_V5_QUEUE', 'collections'),
             'health_batch_size' => (int) env('COLLECTIONS_V5_HEALTH_BATCH_SIZE', 40),
             'recommendation_batch_size' => (int) env('COLLECTIONS_V5_RECOMMENDATION_BATCH_SIZE', 40),

config/forum_security.php

@@ -15,6 +15,8 @@ return [
     'queues' => [
         'moderation' => env('FORUM_SECURITY_MODERATION_QUEUE', 'forum-moderation'),
         'firewall' => env('FORUM_SECURITY_FIREWALL_QUEUE', 'forum-security'),
+        // Plugin jobs have no Horizon consumer. Pause enqueue until a worker exists.
+        'dispatch_enabled' => (bool) env('FORUM_QUEUE_DISPATCH_ENABLED', false),
     ],
 
     'firewall' => [

config/horizon.php

@@ -99,6 +99,7 @@ return [
     'waits' => [
         'redis:broadcasts' => 15,
         'redis:default' => 60,
+        'redis:mail' => 60,
         'redis:notifications' => 90,
     ],
 
@@ -209,9 +210,9 @@ return [
             'maxJobs' => 0,
             'memory' => 128,
             'tries' => 1,
-            // Long-running recommendation rebuild jobs declare timeouts up to 900s.
-            // Keep the worker timeout above that ceiling so Horizon does not kill
-            // healthy jobs before Laravel can enforce the job-level timeout.
+            // Must stay below config/queue.php redis.retry_after (default 1080).
+            // Nightly recommendation rebuilds are chunked; keep headroom for
+            // remaining long jobs on the default queue.
             'timeout' => 960,
             'nice' => 0,
         ],
@@ -228,6 +229,21 @@ return [
             'timeout' => 90,
             'nice' => 0,
         ],
+        // Mail is intentionally isolated: SMTP latency must not block
+        // search/recommendation workers, and mail jobs declare tries=3–5.
+        'supervisor-mail' => [
+            'connection' => 'redis',
+            'queue' => ['mail'],
+            'balance' => 'simple',
+            'autoScalingStrategy' => 'time',
+            'maxProcesses' => 1,
+            'maxTime' => 0,
+            'maxJobs' => 0,
+            'memory' => 128,
+            'tries' => 5,
+            'timeout' => 90,
+            'nice' => 0,
+        ],
     ],
 
     'environments' => [
@@ -242,6 +258,9 @@ return [
                 'balanceMaxShift' => 1,
                 'balanceCooldown' => 3,
             ],
+            'supervisor-mail' => [
+                'maxProcesses' => 2,
+            ],
         ],
 
         'local' => [
@@ -251,6 +270,9 @@ return [
             'supervisor-messaging' => [
                 'maxProcesses' => 2,
             ],
+            'supervisor-mail' => [
+                'maxProcesses' => 1,
+            ],
         ],
     ],
 

config/http_observability.php

@@ -0,0 +1,14 @@
+<?php
+
+declare(strict_types=1);
+
+return [
+    'slow_request' => [
+        'enabled' => (bool) env('HTTP_SLOW_REQUEST_LOG_ENABLED', false),
+        'threshold_ms' => (int) env('HTTP_SLOW_REQUEST_MS', 750),
+        /*
+         * Testing only. When set, replaces measured duration so tests do not sleep.
+         */
+        'test_duration_ms' => env('HTTP_SLOW_REQUEST_TEST_DURATION_MS'),
+    ],
+];

config/logging.php

@@ -123,6 +123,14 @@ return [
             'handler' => NullHandler::class,
         ],
 
+        'slow-http' => [
+            'driver' => 'daily',
+            'path' => storage_path('logs/slow-http.log'),
+            'level' => 'info',
+            'days' => (int) env('HTTP_SLOW_REQUEST_LOG_DAYS', 14),
+            'replace_placeholders' => false,
+        ],
+
         'emergency' => [
             'path' => storage_path('logs/laravel.log'),
         ],

config/metrics.php

@@ -0,0 +1,20 @@
+<?php
+
+declare(strict_types=1);
+
+/**
+ * Hourly artwork metric snapshots (heat, rising, windowed leaderboards).
+ *
+ * Default retention is 30 days so monthly leaderboards and Studio 30d KPIs
+ * have a complete window after warm-up. Override with env; never hardcode.
+ * Studio 30d KPIs use MAX-MIN of cumulative counters, never SUM.
+ */
+return [
+
+    'hourly_snapshot_retention_days' => (int) env('ARTWORK_METRIC_HOURLY_RETENTION_DAYS', 30),
+
+    'hourly_snapshot_prune_chunk' => (int) env('ARTWORK_METRIC_HOURLY_PRUNE_CHUNK', 5000),
+
+    'hourly_snapshot_prune_sleep_ms' => (int) env('ARTWORK_METRIC_HOURLY_PRUNE_SLEEP_MS', 50),
+
+];

config/queue.php

@@ -68,7 +68,10 @@ return [
             'driver' => 'redis',
             'connection' => env('REDIS_QUEUE_CONNECTION', 'default'),
             'queue' => env('REDIS_QUEUE', 'default'),
-            'retry_after' => (int) env('REDIS_QUEUE_RETRY_AFTER', 90),
+            // Must exceed the longest Horizon worker timeout (960s) plus a buffer.
+            // If retry_after is shorter than a running job, Redis re-queues it as a
+            // new attempt and Horizon (--tries=1) marks MaxAttemptsExceeded.
+            'retry_after' => (int) env('REDIS_QUEUE_RETRY_AFTER', 1080),
             'block_for' => null,
             'after_commit' => false,
         ],

config/ranking.php

@@ -77,6 +77,11 @@ return [
         'candidate_pool' => 200,  // top N candidates to run diversity filter on
     ],
 
+    // Unique lock for RankBuildScopeListsJob. Must exceed queue wait + job
+    // timeout (300s). Production wait has been ~2.5h; 6h prevents hourly
+    // stacking while still recovering a dead lock the same day.
+    'scope_job_unique_for' => (int) env('RANK_SCOPE_JOB_UNIQUE_FOR', 21600),
+
     // ── Anti-spam / burst-view damping ──────────────────────────────────────
     'spam' => [
         'views_24h_threshold'     => 2000,

config/sitemaps.php

@@ -54,6 +54,9 @@ return [
         'forum-threads' => [
             'size' => (int) env('SITEMAPS_SHARD_FORUM_THREADS_SIZE', 10000),
         ],
+        'tags' => [
+            'size' => (int) env('SITEMAPS_SHARD_TAGS_SIZE', 25000),
+        ],
         'collections' => [
             'size' => (int) env('SITEMAPS_SHARD_COLLECTIONS_SIZE', 10000),
         ],

config/skinbase_ai_moderation.php

@@ -7,6 +7,7 @@ return [
 
     'queue' => [
         'name' => env('SKINBASE_AI_MODERATION_QUEUE', 'forum-moderation'),
+        'dispatch_enabled' => (bool) env('FORUM_QUEUE_DISPATCH_ENABLED', false),
     ],
 
     'preflight' => [

config/traffic.php

@@ -0,0 +1,15 @@
+<?php
+
+declare(strict_types=1);
+
+return [
+    'online_visitors' => [
+        // Cap SSCAN/all() so moderation pages never SMEMBERS the 2.4M-member index.
+        'index_read_limit' => (int) env('ONLINE_VISITOR_INDEX_READ_LIMIT', 2000),
+        'prune_enabled' => (bool) env('ONLINE_VISITOR_INDEX_PRUNE_ENABLED', false),
+        'prune_scan_count' => (int) env('ONLINE_VISITOR_INDEX_PRUNE_SCAN_COUNT', 500),
+        'prune_max_batches' => (int) env('ONLINE_VISITOR_INDEX_PRUNE_MAX_BATCHES', 200),
+        'prune_max_members' => (int) env('ONLINE_VISITOR_INDEX_PRUNE_MAX_MEMBERS', 100000),
+        'prune_sleep_ms' => (int) env('ONLINE_VISITOR_INDEX_PRUNE_SLEEP_MS', 25),
+    ],
+];