Opening up our π¨ LLM downtime monitoring tool to the community:
We started to maintain this because we were tired of running into API errors without warning. Figured it could help other developers stay in the loop on API performance too.
Hey PH Community,
Tired of unexpected OpenAI outages derailing your production apps? We've got your back. πͺ
We built an internal tool to monitor API performance across OpenAI, Anthropic, Anyscale - and now we're opening it up to everyone. π
Our tool tracks:
β Real-time downtime alerts for key models like GPT-3.5, GPT-4, embeddings
β Live API latency metrics so you can see when models slow down
β Proactive pings every 3 mins so we catch issues before you do
Please let us know what you think!
Well this is going to save me a whole lot of time in the future! An elegant fix to an issue I have been facing and can only assume has been causing just as much grief for other AI system creators. Great job well done!!
Portkey