If Relay connect failure happened on client app side without informative helpful info in local logs to identify the issues. Clients can check the logs from brainCloud portal monitoring for further investigation.
For instance, if client app sends data packets size beyond the limitation of maximum packet size, which is 1024 bytes for relay server, they might get connection broken right away and with local log shows as (this example uses UDP connection):
Relay connect failure: Timed out. Too many packet drops.
Relay connect failure: UDP Receive Failed
I/System.out: RELAY UDP: Timed out. Too many packet drops
Then, if they check from brainCloud portal page -- Monitoring | Recent Errors, they would find the detail of the problem, which cause this issue shows as below, from Error Details | Context box.
Note: Check the info checkbox from Filter by, and click Refresh button, then scroll down to your accordingly log, click to open.