The connector can continue streaming from a secondary node promoted to primary after a fail-over. Would be great to demo this and show the required configuration. Note there already is an example for MySQL.