Between May 23, 22:00 and May 24, 02:30hs (UTC-3) there was an internal error in a backoffice instance, which caused the system to respond with HTTP 502 error in approximately 5% of the received requests during that time frame. The instance was taken out of service and the problem was resolved.