Arguably one dose per year. Arguably several in close sequence, but still very intermittently, such as once a year. Senescent cells accumulate slowly, so there is no reason to dose frequently on an ongoing basis. Different mouse studies of dasatinib + quercetin used single doses or a few monthly doses in sequence. This is true of dasatinib, but varies for other chemotherapeutic senolytics. Bcl-2 inhibitor chemotherapeutics generally want a week or a few weeks of dosing, rather than a single dose, which is a compelling reason to avoid them until better versions come along. The FOXO4-DRI injected peptide was a single treatment of three doses over a week in mice. That again is probably once a year in humans until someone turns up with evidence to suggest otherwise.

Validation of the above educated guess, and some form of dose optimization, will arrive sometime after the introduction of good senescence assays that don't involve staining tissue samples, something that isn't viable for everyday human medicine.

As to what "partially reversed" means, the answer is "read the literature". It is all out there. Search PubMed for "senolytic". Use Sci-Hub to read papers. The comparison pictures of treated and untreated mice in papers that have them are particularly telling, but no substitute for the raw data on arthritis, cardiovascular function, inflammation, neural function, tissue elasticity, etc.

For side effects, a single dose dasatinib pharmacokinetics study in volunteers with lots of references to start you off: https://doi.org/10.1124/dmd.107.018267