For "Maximum document length", you are talking about the Solr connector, correct? In that case it is the maximum size of extracted content that will be sent to Solr. (The connector assumes that when you aren't using the /update/extract handler you are extracting the content upstream of the Solr connection.)
For "Max file size (bytes)", I assume it's the Document Filter transformer you are talking about. This is designed to remove documents whose binary is in excess of the desired limits. It could be used just prior to the Solr connection to remove documents whose extracted content is too large, but we chose to integrate the limit right in the Solr connector because that we can enforce the fact that a user really must have a limit for /update handler content, since it all lives in memory.