1. This is measure of sequencing depth ("library" is different; it refers to cDNAs not reads)
2. No; this is just a way to filter out low-count genes; has nothing to do with sequencing depth or coverage. Also, the number of reads that align to a particular gene isn't very informative because highly-expressed genes and genes with longer transcripts will obviously have more reads aligned to them.

Also, coverage (distinct from sequencing depth) estimates how many times each base in a genome or transcriptome is sequenced, on average. It's not a useful metric for RNA-seq because when a gene is highly expressed, it's going to have more sequencing reads than a lowly-expressed gene.