LLM call Latency increased due to agentops

[rupav]

🐛 Bug Report

🔎 Describe the Bug Give a clear and concise description of the bug. I have a fastapi uvicorn server which serves multiple concurrent requests. In each of the call, I am using LLM call. To monitor the same, I am creaing new agentops session, and then patch then initiate_chat method of the autogen's first most call. But using the same, LLM call latency have been increased by 5x.

🔄 Reproduction Steps

• Init agentops before any LLM calls (on startup)
• create a session on even API call to server (which has internal autogen agents integrated)
• use created session.patch method on initiate_chat method of autogen
• end the session before serving API response to client.

🙁 Expected Behavior Describe what you expected to happen.

• Latency should be minimised

🔍 Additional Context Provide any other context about the problem here. Python: 3.11 agentops: 0.3.6 pyautogen: 0.2.32

[HowieG]

Hey @rupav, thanks for reporting. We are discussing this internally and think we might just need to deploy a server in India to bring down the latency. Will report back when we arrive at a decision thanks

[dot-agi]

Hey @rupav, thanks for reporting. We are discussing this internally and think we might just need to deploy a server in India to bring down the latency. Will report back when we arrive at a decision thanks

Cheap ones available in New Delhi :)