fix feedback

api/.env.example

@@ -530,7 +530,7 @@ LOG_FORMAT=%(asctime)s,%(msecs)d %(levelname)-2s [%(filename)s:%(lineno)d] %(req
 
 # Indexing configuration
 INDEXING_MAX_SEGMENTATION_TOKENS_LENGTH=4000
-# Maximum number of worker threads used for high-quality dataset indexing.
+# Maximum number of worker threads used for high-quality dataset indexing (1-10).
 # Lower this value to reduce memory usage and avoid OOM freezes during re-indexing.
 INDEXING_MAX_WORKERS=2
 

api/configs/feature/__init__.py

@@ -1109,7 +1109,7 @@ class IndexingConfig(BaseSettings):
     )
 
     INDEXING_MAX_WORKERS: PositiveInt = Field(
-        description="Maximum number of worker threads used for high-quality dataset indexing",
+        description="Maximum number of worker threads used for high-quality dataset indexing (1-10)",
         default=2,
     )
 

api/core/indexing_runner.py

@@ -604,7 +604,7 @@ class IndexingRunner:
 
         # High-quality indexing is memory intensive (embedding generation + vector writes).
         # Running too many chunks in parallel can trigger OOM and freeze the service until reboot.
-        max_workers = max(1, min(10, int(dify_config.INDEXING_MAX_WORKERS)))
+        max_workers = max(1, int(dify_config.INDEXING_MAX_WORKERS))
         max_workers = min(max_workers, len(documents)) if documents else 1
         if dataset.indexing_technique == IndexTechniqueType.HIGH_QUALITY:
             with concurrent.futures.ThreadPoolExecutor(max_workers=max_workers) as executor:

docker/.env.example

@@ -1009,7 +1009,7 @@ SENDGRID_API_KEY=
 # Maximum length of segmentation tokens for indexing
 INDEXING_MAX_SEGMENTATION_TOKENS_LENGTH=4000
 
-# Maximum number of worker threads used for high-quality dataset indexing.
+# Maximum number of worker threads used for high-quality dataset indexing (1-10).
 # Lower this value to reduce memory usage and avoid OOM/freezes during re-indexing.
 INDEXING_MAX_WORKERS=2