I build a RouterQueryEngine pipeline, and test it on 50 questions. It happened at least 1 time per test round. I am using openai API. The WARNING info is as follows:
WARNING:llama_index.llms.openai_utils:Retrying llama_index.llms.openai_utils.completion_with_retry.<locals>._completion_with_retry in 4.0 seconds as it raised Timeout: Request timed out: HTTPSConnectionPool(host='api.openai.com', port=443): Read timed out. (read timeout=600).
Retrying llama_index.llms.openai_utils.completion_with_retry.<locals>._completion_with_retry in 4.0 seconds as it raised Timeout: Request timed out: HTTPSConnectionPool(host='api.openai.com', port=443): Read timed out. (read timeout=600).
