1) for the most dangerous of diseases, we tend to know their mechanism (gene(s) responsible), precisely because if they're dangerous then they are interesting genes and have been studied in the lab already. That's not to say we know everything - but that information about genes is currently not limiting (for the field).
2) But what you can do in a close environment of the lab with clean rooms, infinite cells, lots of time, infinite do-overs, 30% success rates, and single-cell type environments you can't do in a live body. We've taken apart enough 'cars' to know how the engine mostly works - which are the ignorable parts and which are critical. We've even practiced remaking certain parts, even being creative about making better parts (synthetic bio) to upgrade the engine - however - doing all of the above on a running car that's going 60mph is a whole different story. Currently, the largest challenge is delivering your genetic payload to exactly, and only exactly where you want it. Nowhere more, nowhere less, nowhere wrong, and just right. That delivery is key (and part of why Cas9/crispr is a big deal, it solves part of the problem (where in a genome)). But even if we can target where in a genome, we still need to target where in an organism, and where not in an organism. Delivery is the current limitation. You are made of 3 billion base pairs of information duplicated between a few trillion cells; A 0.00001% mis-delivery rate of this absolutely stunning, perfectly designed new gene/part/function is likely unacceptable.
3) Aging is a lot of things depending on who you ask. Generally mammalian cells are designed to stop growing - otherwise you'd be a thousand-pound sphere of goop at this point rather than a well-defined human with shape. This is a good thing. But it means there are built-in limits on how many times a cell intended to divide. Imagine a book with blank pages up front and in back, where the copy machine can't copy the covers, so it always skips the first and last page. Human cells have about 60 blank pages before they start eating into the text (genetic code) at which point they hit their last life and just try to never die (senesce). There are things we can do to add more pages with genetic therapy, but you got to be careful you don't cause rampant growth (cancer). SO aging is tricky, and a huge frontier that we know very very little about (hard to do experiments where your mean time to finish the assay is 60 years). But there's nothing inherently intractable about the concept that gene-therapies shouldn't be able to affect.
Small molecules ('drugs' as you think of them), are great at killing invaders - things that are distinctly non-human (bacteria/viruses/fungi/etc.). They are not good at affecting 'disorders', cancers or other issues where your own body is over/under/mis-reacting. For that you need to change/alter/upgrade the body's own toolkit. This ability is what synthetic biology promises - access to that toolkit.