[Question]: Still unable to restrict rate limit when generating knowledge graph

[Joey233qwq]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (Language Policy).
• Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) (Language Policy).
• Please do not modify this template :) and fill in all the required fields.

Describe your problem

I have updated to nightly version, and also add "MAX_CONCURRENT_CHATS=10" in my docker/.env ( or should I add "export MAX_CONCURRENT_CHATS=10" instead ?). However, when extracting, there are still so many tokens consumed in 1 minute：

...
22:02:11 Page(1164): Embedding chunks (40.25s)
22:04:00 Page(1164): Indexing done (108.99s). Task done (573.69s)
22:06:49 Entities extraction of chunk 368 277/800 done, 6 nodes, 0 edges, 8425 tokens.
22:06:49 Entities extraction of chunk 355 278/800 done, 0 nodes, 0 edges, 7128 tokens.
22:06:50 Entities extraction of chunk 358 279/800 done, 9 nodes, 0 edges, 7882 tokens.
22:06:50 Entities extraction of chunk 351 280/800 done, 8 nodes, 0 edges, 7669 tokens.
22:06:50 Entities extraction of chunk 345 281/800 done, 8 nodes, 8 edges, 9349 tokens.
...

[Joey233qwq]

Now I set MAX_CONCURRENT_CHATS=1 instead, and find out only 1 chunk extraction in 1 second. I wonder if MAX_CONCURRENT_CHATS just limit the number of parallel operation when extracting. And if this is true, this setting can't really help to restrict rate of model, since every extraction finishes quickly, and token per minute will still be so high. How can I controll request per minute to restrict the rate instead?

[KevinHuSh]

For now, MAX_CONCURRENT_CHATS is the only way to control it.