Hey, guys. I have recently integrated helicone with our production system. It works well except for one of our features which results in long responses (avg ~40s). I might be wrong but is this due to some response time/log threshold set internally by helicone? and is it changeable through a header?
Hey. So we stopped using proxy and replaced it with helicone's logger from the sdk to integrate with langchain callbacks. It was a tad bit more effort but it is working well for us π