Retryable errors surfacing from transaction functions

[2hdddg]

Retryable errors that happens in a cluster are not retried and transactions isn't committed and error returned to the user application.
These kind of errors are returned to application during leader election:
Connection error: dial tcp: i/o timeout
Connection error: dial tcp 34.78.37.125:7687: connect: connection refused
Neo.ClientError.Cluster.NotALeader