CODEWITHSUNDEEP
azure-cognitive-servicesazure-qna-maker
Alex
Alex

Reputation: 210

Handle throttling and/or quota in qna maker

How can you handle throttling and/or quota in qna maker? I'm sure that I'm not the only one who has this question, and maybe, you'll give me some ideas.

I have a chatbot that handles specific problems, but the user can also type free text, so I want to be able to respond to some social questions like hi, hello, bye bye, thank you or swearing. So QnA seams perfect for this.

The problem is that 10k requests quota means aprox 300 requests per day. Presuming 30 chats per day with an average handling time of 10-15 minutes with lots of questions, I think the quota is exceeded too soon.

Any ideas?

Upvotes: 2

Views: 181

Answers (1)

Sebastian Zolg
Sebastian Zolg

Reputation: 1961

QnA Maker is now GA since May 7, 2018. With GA the architecture of QnA Maker changed.

QnA Maker announced General Availability on May 7, 2018 at the \build\ conference. QnA Maker GA has a new architecture built on Azure. Knowledge bases created with QnA Maker Free Preview will need to be migrated to QnA Maker GA. QnA Maker Preview will be deprecated in November 2018. For more information about the changes in QnA Maker GA, see the QnA Maker GA announcement blog post.

With GA there are many pricing tiers for QnA Management and the underlaying azure search service. Non of those tiers have limitations on request, so this question is no longer valid.

In case future readers want to overcome other scaling limitations, there is a documentation on how to build a scaled QnA Maker infrastructure.

See https://learn.microsoft.com/en-us/azure/cognitive-services/qnamaker/how-to/business-continuity-plan.

This is for business continuity but could easily be adapted to scaling scenarios as well.

Upvotes: 1

Related Questions