[Mailer] Improve documentation for mailer failover and round-robin transports

[pableu]

I think the current documentation for the failover transport is not correct.

If the sending fails, the mailer won’t retry it with the other transports

This isn't what really happens. In my tests (and from studying the code), it seems that the mailer simply retries with the next available transport.

This is also what one would expect from a failover. Failing the first time something goes wrong isn't really a failover ;-)

So the only difference between failover and round-robin is that the latter selects a transport at random. My proposed change reflects that more clearly.

[pableu]

Updated. Let me know what you think.

[OskarStark]

For me this is more understandable, thank you

[javiereguiluz]

Pablo, thanks for improving this article. Things are now much easier to understand!