@Pav do you have a tenant I can look at?
One thing I want to clarify for any one finding this thread is that moving your calls over from legacy keys to one of the supported authentication mechanisms is putting the calls through our API Proxy. The endpoints themselves are exactly the same. It is entirely possible that errors like this were failing silently before (connection issues like timeouts, resets, etc). Now Our API Proxy sees these and returns a 503 which is the proper response code for requests to downstream services failing. So if these increases correspond with moving off legacy keys it might this difference.
@bradb, checking through the logs, I found the following number of times the issue has been reported for the tenant.
- May: 57
- June: 21
- July: 12
This was using the Legacy API Key under the original API Endpoint address.
Although, I am not too concerned given that 1000s of API Calls are occurring during the month, and the error has not been affecting things.