Hello dev community,
We call the public/users/discover endpoint from our server as a mechanism to share state. We've been using it for years without any issues.
Starting on May 25, 2021 we started receiving 503
s from the /database/1/<container>/production/public/users/discover
endpoint for a high-percentage of our calls.
Sample Response
Status Code: 503
Response Body:
{
"uuid" : "e8df63ea-14f5-4e5c-9444-68cc3213cda8",
"retryAfter" : 173,
"serverErrorCode" : "THROTTLED"
}
I've read that there is a request limit of 40 requests/second, but we're nowhere near that volume. I'd estimate we're closer to 0.3 requests/second. Yet, 90+% of our calls return that response with a 503
status code.
Documentation indicates that a throttled call should respond with a 429
status code, so the 503
is confusing me. Plus the retryAfter
response parameter never changes. It's always 173
, so I can't even write smart retry logic to wait for an appropriate time to call again.
I've reviewed our activity in the Telemetry and Logs sections of the CloudKit console, and they confirm that we're not close to a throttled threshold.
Has anyone else experienced this?
Thanks in advance for any advice!