perf(ai-moderation): speed up analysis queue (ramai + sepi)

- Parallelize per-user reputation/profile fetches in textBatchProcessor
  (was a serial ~2N DB/Redis round-trip loop per sub-batch; now Promise.all
  over unique users). Cuts per-batch latency, biggest win on small/quiet
  batches.
- Make the LLM concurrency semaphore dynamic (cached per config value) instead
  of frozen at import time, so AI_LLM_MAX_CONCURRENT is tunable without code
  change and reflects current config.
- Bump AI_LLM_MAX_CONCURRENT default 5 -> 8 (gemini-flash-lite is cheap; helps
  throughput when busy).
- Lower AI_ANALYSIS_DEBOUNCE_MS 500 -> 250 (snappier first-message analysis
  when quiet).
- Lower AI_ANALYSIS_RECOVERY_INTERVAL_MS 15000 -> 10000 (stuck/errored
  messages re-analyze sooner).

tsc, biome, vitest (129) all clean.

services/discord-gateway/src/modules/ai-moderation/llmClient.ts

@@ -18,7 +18,20 @@ const log = createChildLogger("llm-client");
 // Concurrency limiter for LLM API calls (inlined from concurrencyLimiter.ts)
 // ---------------------------------------------------------------------------
 
-const llmSemaphore = pLimit(config.AI_LLM_MAX_CONCURRENT ?? 5);
+// The limiter is cached per configured concurrency value so it can be tuned
+// (env / BWS) without a code change and always reflects the current config —
+// a module-level `pLimit(config.X)` would freeze the cap at import time.
+let llmSemaphore = pLimit(config.AI_LLM_MAX_CONCURRENT ?? 5);
+let llmSemaphoreLimit = config.AI_LLM_MAX_CONCURRENT ?? 5;
+
+function getLlmSemaphore() {
+  const wanted = config.AI_LLM_MAX_CONCURRENT ?? 5;
+  if (wanted !== llmSemaphoreLimit) {
+    llmSemaphore = pLimit(wanted);
+    llmSemaphoreLimit = wanted;
+  }
+  return llmSemaphore;
+}
 
 let activeCount = 0;
 let pendingCount = 0;
@@ -30,7 +43,7 @@ export async function withLlmConcurrency<T>(fn: () => Promise<T>): Promise<T> {
     "Queuing LLM request",
   );
 
-  return llmSemaphore(async () => {
+  return getLlmSemaphore()(async () => {
     pendingCount--;
     activeCount++;
 

services/discord-gateway/src/modules/ai-moderation/textBatchProcessor.ts

@@ -214,20 +214,28 @@ export async function runTextOnlyBatch(
         asOf?: number | null;
       }
     >();
-    for (const msg of batch) {
-      if (!userContexts.has(msg.user_id)) {
-        const rep = await initializeUserReputation(msg.user_id, msg.guild_id);
-        const repAttrs = formatReputationAttrs(rep);
-        const repXml = `<user_reputation ${repAttrs}/>`;
-        userContexts.set(msg.user_id, repXml);
-      }
-      if (!userProfiles.has(msg.user_id)) {
-        const profile = await getUserProfile(msg.user_id);
-        userProfiles.set(msg.user_id, {
-          text: profile?.profile_summary ?? "",
-          asOf: profile?.last_analyzed_at ?? null,
-        });
-      }
+    // ── Per-user reputation + profile context (fetched ONCE per unique user,
+    //    in parallel — was a serial per-message loop that cost ~2N sequential
+    //    DB/Redis round-trips per sub-batch and dominated latency on small
+    //    batches). ─────────────────────────────────────────────────────────
+    const uniqueUserIds = [...new Set(batch.map((m) => m.user_id))];
+    const batchGuildId = batch[0]?.guild_id ?? "";
+    const userFetches = await Promise.all(
+      uniqueUserIds.map(async (uid) => {
+        const [rep, profile] = await Promise.all([
+          initializeUserReputation(uid, batchGuildId),
+          getUserProfile(uid),
+        ]);
+        return { uid, rep, profile };
+      }),
+    );
+    for (const { uid, rep, profile } of userFetches) {
+      const repAttrs = formatReputationAttrs(rep);
+      userContexts.set(uid, `<user_reputation ${repAttrs}/>`);
+      userProfiles.set(uid, {
+        text: profile?.profile_summary ?? "",
+        asOf: profile?.last_analyzed_at ?? null,
+      });
     }
     const userProfilesBlock = buildUserProfilesBlock(userProfiles);
 

services/discord-gateway/src/shared/config/index.ts

@@ -177,7 +177,7 @@ export const configSchema = z
     QDRANT_URL: z.string().optional(),
     QDRANT_COLLECTION: z.string().default("gmw_text_moderation"),
     QDRANT_API_KEY: z.string().optional(),
-    AI_LLM_MAX_CONCURRENT: z.coerce.number().int().positive().default(5),
+    AI_LLM_MAX_CONCURRENT: z.coerce.number().int().positive().default(8),
     AI_LLM_IMAGE_MAX_DIMENSION: z.coerce
       .number()
       .int()
@@ -226,11 +226,11 @@ export const configSchema = z
       .default(5),
 
     // ── AI Analysis Timing ──────────────────────────────────────────────
-    AI_ANALYSIS_DEBOUNCE_MS: z.coerce.number().positive().default(500),
+    AI_ANALYSIS_DEBOUNCE_MS: z.coerce.number().positive().default(250),
     AI_ANALYSIS_RECOVERY_INTERVAL_MS: z.coerce
       .number()
       .positive()
-      .default(15000),
+      .default(10000),
     AI_ANALYSIS_ERROR_COOLDOWN_MS: z.coerce.number().positive().default(30000),
 
     // ── AI Analysis Batch ───────────────────────────────────────────────