1. Correct, besides vCenter Server events, tasks, and alarms and ESXi 4.x and greater you must configure the source directly.
2. This depends on what you are trying to achieve. You can filter based on source or hostname (could be different depending on if you are using an aggregation server) or you could filter by a pattern in the log messages if you can determine one for the device.
"These sources will not appear in Log Insight configuration as it's not configured from Log Insight." - I am not sure what you mean here, but the sources will appear in the interactive analytics page by default.
"I was expecting the configuration to be done from Log Insight so we can control what's coming." - Two things here, first there are way too many sources to support and in general syslog servers today do not typically configure sources for you. Second, you say you want to control whats coming - I assume you mean control what logs messages are and are not sent from the source? If so, this is not possible today, but I would ask why do you want to do that? In general, it is a best practice to send all of your logs as you may need them all to perform a RCA.