Description
When creating connection instance using the Failover transport the time to create a connection can be substantially longer than when using raw TcpTransport instances. Problems exists in the failover reconnect method and oneway send method that cause the initial send call for ConnectionInfo to pause for up to 1 second. Also the way that the InactivityMonitor configures its initial activity delay check results in the delay always being zero unless one has been set on the broker side so an initial check is always done which gets in the way of the normal connection process.