When will logs get dropped by this plugin?

Hi,
This plugin looks really interesting, I have a couple of questions regarding particular lines of code:

sumologic-docker-logging-driver/logger.go

Line 115 in 7ef84d5

    
           logrus.Error(fmt.Errorf("%s: Log batch queue full, dropping oldest batch", pluginName))

sumologic-docker-logging-driver/logger.go

Line 91 in 7ef84d5

    
           logrus.Warn(fmt.Sprintf("%s: Log is too large to batch, dropping log. log-size: %d bytes",

These lines seem to indicate circumstances in which logs will be lost (log too large to batch and batch full). Could you clarify the circumstances in which this would occur (not sure what exactly 'batch' is or what size it may be)?

Is anyone using this plugin in production?

Cheers!
Bryan

There is a memory queue pile the message before sending and there is a parameter sumo-queue-size for tuning how large of data it can hold