Suggestion on billing/ system quota of AI-relative GCP APIs

Google Cloud Build with AI AI APIs
,
1

I just encountered trouble with billing of Videointelligence AI.
Luckily GCP billing support has been very helpful and efficient to help me solving the issue !

Here is what happened and what I would like to suggest that may improve the service.

  1. I have some videos need Text Detection/ OCR, using pure REST API is hard to automate the process,

so instead I used the Python Client Wrapper to submit requests.

  1. Unfortunately because of a script bug, I submit the wrong batch from buckets, which takes a ‘*’ wildcard in a folder.

  2. I instantly realized my mistake then manually make cancel API requests to Operations. I checked the Operations status it shows me it’s USER CANCELLED

  3. The billing/pricing did not show up immediately until the next day I checked it.

The total minutes still includes those Operations got cancelled.

  1. I contacted the Billing support, It has been a great help to revoke the billing.

Summary of the problem:

  1. AI relative APIs are expensive, users do care about the usage and do not what to waste money on mistakes.
  2. It’s indeed my mistake submitting the wrong batch, but the API seems error prune.

A. There is no system quota/ limits setting on how much total you can spent, you might spend a lot for simple mistake. Human mistakes should be minimized by simple fool proof mechanics.

B. Operation cancellation functionality is questionable, from what I experienced.

C. The console panel does not have the functionality to monitor and control usage, neither the Client Wrapper does(you cannot retrieve Operation names/ Ids).

Suggestions:

  1. Again, a total limit/ cap is important to avoid loss.

  2. Credit based allocation might be great, you only buy credits you needs.

User could choose auto replenish or not.

  1. Console/ Panel/ Library experience could be improved. Some other cloud services has interfaces to try and test.

  2. Cancellation functionality might need inspection.

Thanks.

2

Hi @virtualdruid70 ,

Welcome to Google Cloud Community!

I appreciate your suggestions. You may log this as a Feature request here. Also, this way you could have visibility on the progress of your request as it is publicly available. Please note that I can’t provide any details or timelines at this moment. However, you may keep an eye on the release notes for any latest updates or new features related to Cloud Vision API.

1 Like