I imagine that, at this stage, rather than the engineering stage, the cycle rate is arbitrary.
Although the ultimate rate woould be limited by the evolution lifetime of the discharge into plasmoid, it is more likely to be a practical consideration about how fast we can remove waste heat, or how fast capacitors can be charged, discharged and switched. As stated by vansig the minimum rate would need to facilitate efficient maintenance of chamber and contents temperature for initiating the next discharge. There have also been suggestions on this site about a natural resonant rate supported by the electrical parameters, and a flat-out rate without switches where the next discharge can start as soon as sufficient input voltage has built.