Thank you Matt for your reply however, none of your scenarios seem to fit.
Both servers purge the BINLOG after seven(7) days. Both servers are configured identically EXCEPT for 'auto-increment-offset'. Yes, both servers are suppose to have identical data sets. BTW, this problem occurs on both master servers.
You are correct about the error message. The Slave is trying to connect to mysql02 and then gets the error.
The binlog hasn't been purged. The binlog on mysql02 is just missing the GTID (excample: 819c985c-d384-11e3-a621-00259002979a:154866). However, the relay-log has the GTID - why is that!? And doing a SHOW SLAVE STATUS the RETRIEVED_GTID_SET shows the GTID was retrieved.
The point is WHY is a GTID missing from the EXECUTED_GTID_SET. There is no log message as to why.
What I have noticed is that a skip occurs occasionally when the servers are restarted. This has been happening for sometime and I opened a MySQL bug years ago about it:
https://bugs.mysql.com/bug.php?id=74687