From f4f8fc051e10bf4d5396bcbdd36a5f7ee6db68e5 Mon Sep 17 00:00:00 2001
From: Waleed <walif6@gmail.com>
Date: Fri, 29 Aug 2025 21:09:09 -0700
Subject: [PATCH] improvement(kb): add fallbacks for kb configs (#1199)

---
 apps/sim/background/knowledge-processing.ts | 12 +++++------
 apps/sim/lib/knowledge/documents/service.ts | 24 ++++++++++-----------
 2 files changed, 18 insertions(+), 18 deletions(-)

diff --git a/apps/sim/background/knowledge-processing.ts b/apps/sim/background/knowledge-processing.ts
index 0260ed096..02c4108ad 100644
--- a/apps/sim/background/knowledge-processing.ts
+++ b/apps/sim/background/knowledge-processing.ts
@@ -26,15 +26,15 @@ export type DocumentProcessingPayload = {
 
 export const processDocument = task({
   id: 'knowledge-process-document',
-  maxDuration: env.KB_CONFIG_MAX_DURATION,
+  maxDuration: env.KB_CONFIG_MAX_DURATION || 300,
   retry: {
-    maxAttempts: env.KB_CONFIG_MAX_ATTEMPTS,
-    factor: env.KB_CONFIG_RETRY_FACTOR,
-    minTimeoutInMs: env.KB_CONFIG_MIN_TIMEOUT,
-    maxTimeoutInMs: env.KB_CONFIG_MAX_TIMEOUT,
+    maxAttempts: env.KB_CONFIG_MAX_ATTEMPTS || 3,
+    factor: env.KB_CONFIG_RETRY_FACTOR || 2,
+    minTimeoutInMs: env.KB_CONFIG_MIN_TIMEOUT || 1000,
+    maxTimeoutInMs: env.KB_CONFIG_MAX_TIMEOUT || 10000,
   },
   queue: {
-    concurrencyLimit: env.KB_CONFIG_CONCURRENCY_LIMIT,
+    concurrencyLimit: env.KB_CONFIG_CONCURRENCY_LIMIT || 20,
     name: 'document-processing-queue',
   },
   run: async (payload: DocumentProcessingPayload) => {
diff --git a/apps/sim/lib/knowledge/documents/service.ts b/apps/sim/lib/knowledge/documents/service.ts
index 3654e541e..4d8c4c478 100644
--- a/apps/sim/lib/knowledge/documents/service.ts
+++ b/apps/sim/lib/knowledge/documents/service.ts
@@ -17,8 +17,8 @@ import type { DocumentSortField, SortOrder } from './types'
 const logger = createLogger('DocumentService')
 
 const TIMEOUTS = {
-  OVERALL_PROCESSING: env.KB_CONFIG_MAX_DURATION * 1000,
-  EMBEDDINGS_API: env.KB_CONFIG_MAX_TIMEOUT * 18,
+  OVERALL_PROCESSING: (env.KB_CONFIG_MAX_DURATION || 300) * 1000,
+  EMBEDDINGS_API: (env.KB_CONFIG_MAX_TIMEOUT || 10000) * 18,
 } as const
 
 /**
@@ -38,17 +38,17 @@ function withTimeout<T>(
 }
 
 const PROCESSING_CONFIG = {
-  maxConcurrentDocuments: Math.max(1, Math.floor(env.KB_CONFIG_CONCURRENCY_LIMIT / 5)) || 4,
-  batchSize: Math.max(1, Math.floor(env.KB_CONFIG_BATCH_SIZE / 2)) || 10,
-  delayBetweenBatches: env.KB_CONFIG_DELAY_BETWEEN_BATCHES * 2,
-  delayBetweenDocuments: env.KB_CONFIG_DELAY_BETWEEN_DOCUMENTS * 2,
+  maxConcurrentDocuments: Math.max(1, Math.floor((env.KB_CONFIG_CONCURRENCY_LIMIT || 20) / 5)) || 4,
+  batchSize: Math.max(1, Math.floor((env.KB_CONFIG_BATCH_SIZE || 20) / 2)) || 10,
+  delayBetweenBatches: (env.KB_CONFIG_DELAY_BETWEEN_BATCHES || 100) * 2,
+  delayBetweenDocuments: (env.KB_CONFIG_DELAY_BETWEEN_DOCUMENTS || 50) * 2,
 }
 
 const REDIS_PROCESSING_CONFIG = {
-  maxConcurrentDocuments: env.KB_CONFIG_CONCURRENCY_LIMIT,
-  batchSize: env.KB_CONFIG_BATCH_SIZE,
-  delayBetweenBatches: env.KB_CONFIG_DELAY_BETWEEN_BATCHES,
-  delayBetweenDocuments: env.KB_CONFIG_DELAY_BETWEEN_DOCUMENTS,
+  maxConcurrentDocuments: env.KB_CONFIG_CONCURRENCY_LIMIT || 20,
+  batchSize: env.KB_CONFIG_BATCH_SIZE || 20,
+  delayBetweenBatches: env.KB_CONFIG_DELAY_BETWEEN_BATCHES || 100,
+  delayBetweenDocuments: env.KB_CONFIG_DELAY_BETWEEN_DOCUMENTS || 50,
 }
 
 let documentQueue: DocumentProcessingQueue | null = null
@@ -59,8 +59,8 @@ export function getDocumentQueue(): DocumentProcessingQueue {
     const config = redisClient ? REDIS_PROCESSING_CONFIG : PROCESSING_CONFIG
     documentQueue = new DocumentProcessingQueue({
       maxConcurrent: config.maxConcurrentDocuments,
-      retryDelay: env.KB_CONFIG_MIN_TIMEOUT,
-      maxRetries: env.KB_CONFIG_MAX_ATTEMPTS,
+      retryDelay: env.KB_CONFIG_MIN_TIMEOUT || 1000,
+      maxRetries: env.KB_CONFIG_MAX_ATTEMPTS || 3,
     })
   }
   return documentQueue