I ask because the errors don't seem to line up with what the documentation says.
We have a daily limit of 1 million requests. Upon reaching about 995,000 requests, we started getting errors about hitting the limit and requests flat-lined. This happened around 7:20pm. The dashboard says "Daily quotas reset at midnight Pacific Time (PT)". However, requests started going through again at about 8:30pm.
Since 995,000 < 1,000,000 and 8:30pm != midnight, this leaves me with the feeling that we can't actually predict and prevent hitting the rate limit, or know when it will reset.
What do I need to know in order to do occasional heavy-use of the API while staying under the limits? Nothing in the documentation or dashboards gives me what I need.