False 429 error on first API query

[Frederic Malenfant]

Some users of our app Bridge24 have that issue:

They need to query informations about a "member", using that endpoint: 

GET https://trello.com/1/members/(idMember)?key=(myKey)&token=(myToken)&fields=fullName,username,status,avatarUrl,memberType

They receive a 429 error in return, "too many requests".

But, in that case, it's the first request!

Also, I get these in the reasponse headers:

X-Rate-Limit-Api-Key-Remaining: 297
X-Rate-Limit-Api-Token-Remaining: 97
X-Rate-Limit-Member-Remaining: 372

How can it be a 429 ?

I think that it occurs for people that are not in a workspace, they only have been invited to other workspaces, but they don't see them in the list.

I think that the good response should be "403 forbidden".

[Frederic Malenfant]

Ok, let me add some details....

After more testing, we really had to do 100 calls to "members", so the 429 message can be ok, but something went wrong.

The X-Rate-Limit-Api-Token-Remaining header is going from 100 to about 25 after, 10 secondes, it is reset to 100.

We clearly see that the response move from 100, 99, 98, ... 25, 100.

But, it looks like the 100 returned is not applied server-side, because 25 query later, we start getting "429". 
So, I suppose that, somewhere, the limit continue to move at 24, 23 , 22... even if the header we get in response was  reset to 100 after the X-Rate-Limit-Api-Token-Interval-Ms limit.

[Lorenz]

I have the same issue on the same endpoint. Although the response headers indicate, that the rate limit has been reset, the API does not act accordingly. Sometimes, I even get 429 on the first call after waiting for several minutes for a reset.

Edit: I guess, I missed that there is an extra limit for /1/members of 100 request per 900 secondes.