OCR/read timeouts keep happening even with simple requests of 1 page PDFs

I'm trying to create a module that automatically reads and interprets PDFs extracted from emails.

Using OCR/read from Document Intelligence it should be able to read and extract the text from the PDFs and afterwards it will be handled by an LLM. However, almost every time I get timeouts (see picutre). It seems like it has something to do with the subscription, but the S0 tier should be able to handle 15 pages per second. Does it have to do with the region?