Six suggestions things that might be of use.
(1) could you switch to Acetonitrile based mobile phase instead of methanol?
(2) perhaps the HPLC is "contaminated" you could try cleaning the HPLC unit itself using the Agilent cleaning solvent, I can give you a protocol if you like ?
(3) you could try using a higher grade MeOH, like something used for a QTOF MS.
(4) are there any other MRMs you could try for PFOS that don't product a signal in the blank, we typically use 2 MRMs per analyte of interest, sometimes 3.
(5) could you try a completely different column chemistry, C18 if not being used already, or Biphenyl, or PFP if using C18 now. Agilent and Phenemonex have good PFP columns.
(6) Could you adjust your gradient to a very fine increase in mobile phase B to try separate out the interfering peak, assuming it's not something in the MeOH itself.
You may have already solved the problem, but I recommend this Agilent application note: https://www.agilent.com/cs/library/applications/5991-7863EN.pdf
PFOS and other PFASs can originate from within the system itself. You can replace a lot of the pump components to reduce the background, but it may be easier to use a delay column after the pump mixer as described in the application note to separate the background peak from your analyte peak.
Hope that helps.