Changing Azure ML Endpoint number of concurrent requests

Question

I'm trying to deploy a few Azure Machine Learning endpoints to production that we can then use for our applications and for other customers to access as samples. Our applications allow users to monitor RTSP cameras. Once they choose to monitor we will process every frame of each stream. I was thinking we could process them by making requests to these endpoints and splitting the traffic between CPU and GPU deployments.

However, I'm noticing that just after having two concurrent requests I start receiving an error 429- Too many requests for the model. I've checked the documentation for this and it seems that I can configure the max concurrent request but I can't seem to change it within the studio. Has anyone tried changing this? Or am I going at it the wrong way ?

Thanks in advance.

Answer 1

@Pedro Brito The default requests per second value should be much more than what you are seeing since you have mentioned only two concurrent requests were passed to your online endpoint. The limits are defined at region level so if you have other deployments that are using up the same you could run into this error. The process to increase the endpoint level quota is through an azure support case, just follow the steps mentioned here and you should be good to go.

If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.