Tolga Gerekci

When editing HPA on a k8s 1.24 cluster, instead of autoscaling/v2 editor shows autoscaling/v2beta1

3

When editing HPA on a k8s 1.24 cluster, editor shows `autoscaling/v2beta1` instead of `autoscaling/v2`. However if I run `kubectl edit` autoscaling/v2 is shown correctly. Big issue is, autoscaling/v2 has some...

Properly handle OpenAI JSON error responses when expecting stream response

2

OpenAI returns `application/json` content-type for errors even when expecting a stream in cases like token limit error or other openai related errors. Right now, `axios-fetch-adapter.js:getResponse` only throws a very generic...