Our lab and our colleagues have done steady state plasma PK studies in mice in a couple different ways - Alzet SC or IP implantable pumps, or by infusion pump cannulated to jugular or femoral vein. We get multiple blood samples per mouse (typically 2-3) via retro orbital plexus bleeds and facial vein bleeds, yielding 25-50 uL of plasma per sample. We appropriately anesthetize the mice for the sampling procedure to minimize pain and stress. We then use highly sensitive LC MS/MS assays to quantitate the drug. If you don't have a MS or adequate sensitivity vs low sample volume by whatever modality you're using, then multiple samples per mouse won't be feasible, and you'll have to do a sacrifice design. Hope this helps.
I haven't done any PK studies so I can't answer that part, but you can easily get 200-300 ul of plasma from a rat retro orbital bleed, depending on how frequently you need to take it (we've done up to weekly). That said, you could take less blood and increase the frequency and have more than mouse blood would render, although you would likely need to deliver more drug to the larger animal.