The serpentinite transformation of ultramafic minerals at depth in the crust (and perhaps a long time ago) is the most likely ultimate cause, and then we're seeing delayed leakage to the surface, perhaps via clathrates, perhaps direct gas seepage.

The chief factor in my thinking is that Nli Fossae IS a well-known volcanic centre and had activity over a long period of time, but relatively little in the last few hundred million years.

The high local concentration of methane and the very short lifetime is a mixing phenomenon - the gas escapes locally, is relatively concentrated for a while, then gets diluted into the atmosphere of Mars (where it has the 300-year lifetime before it's oxidised away).

Unfortunately, this 2-stage lifetime model does not require large volumes of methane or high emission rates, but it IS a sign of ongoing geologocal/chemical processes on mar.