What qualifies as a "request" in Google Video Intelligence API?

[hithesh111]

I have a Requests per Minute quota limit of 120 and Backend time of 600 seconds per minute.
Even when I annotate 20 videos together (distributed using multiple cores on Databricks ), it shows the following error. However, since 20 is far less than 120, I can't seem to identify why it exceeds 120 requests per minute. So I'm wondering if any other thing other than ".annotate()" function could be classified as a request.

429 Quota exceeded for quota metric 'Requests' and limit 'Requests per minute' of service 'videointelligence.googleapis.com' for consumer 'project_number:__________'. [reason: "RATE_LIMIT_EXCEEDED"

[nceniza]

Hi,

I would suggest creating a case for this, as we cannot view logs and your actual code from here (also is not advisable since it may include sensitive details about your project) Specially if you think this a quota/billing mis match. You can try to create a case here[1]

[1]https://cloud.google.com/contact

[kolban]

My tests would likely be to reduce the number of parallel videos to 1 then 2 then 3 and see if new symptoms/patterns show up.   I also think that there are quota pages in the console that report specifically which quotas have been breached.  I'd also look in Cloud Logging to see if there are clues in  there.   I'm also noting that the quota is 120 per minute.  Remember that is 120 per minute for the WHOLE of the GCP project.   Do you perhaps have additional work being run in your project other than the work you yourself are doing?   Maybe don't run any work yourself for 5 minutes and look at the quota usage for your project during that 5 minute period.  Is it zero?