We have changed the way we calculate the available monthly quotas:
Before August 1st 2022, the monthly quotas for processing units (PUs) and requests were calculated in such a way that new PUs/requests were added in an ongoing manner, i.e. for Enterprise-S (400,000 PU per month), one PU was added every 6.7 seconds.
After August 1st 2022, the monthly quotas are reset on every 1st of the month. In the example above for Enterprise-S, the available volume will be reset to 400,000 PU on every 1st of the month.
The reasons for this change:
Based on user feedback, we found that the previous approach was difficult to understand.
What does this mean for you:
(In the short term) If you only had a small quota available on July 31st, you are lucky because you received the full monthly quota overnight.
If you are a monthly quota user, nothing changes for you - you can continue to use the same volume every month, both for PUs and requests. If your available quota becomes very small towards the end of the month, do not worry - it will be reset shortly afterwards.
If you are a fixed-quota user, there are no changes at all.
On a side note - while we are doing this upgrade, the available quotas are currently not visible in the Dashboard. We apologise for this. However, you will still be able to see the ongoing usage in the Usage tab.
what might have happen (and it has indeed happened in your case) is that there was a bug in the past, missing some of the consumption, effectively showing the numbers too high. This issue was now solved and the numbers are correct.
We understand that this might have caused inconvenience to you, which is why we have reached to you personally on how to mitigate it.
We have some code that looks for the rate limiting headers, (retry-after, x-ratelimit-remaining) and it’s upset because those headers seem to have dropped off. Would that be related to these quota changes?