So the company I work for takes care of the electrical in our local school system. We are in the process of converting one of these schools to Sylvania LED bypass tubes from F32T8 fluorescent. One light in particular, dubbed by me as "the lamp killer", is in a high trafficked custodial closet, powered by an Advance REL-4P32-LW-RH-TP, that gets probably 20-30 power cycles daily. No matter what lamps I put in it they all go mercury starved within 6 months. So my patience finally ran out this morning when I went to get LED tubes in this closet, and the new Sylvania lamps I put in less than 6 months ago were all mercury starved. So it proceeded to get the early LED treatment  since I haven't made it to that side of the building yet. There's 2 other fixtures with the same ballasts in the same closet that don't do this, so I wonder if it is a faulty ballast that was constantly overdriving the lamps. These ballasts, especially the 3 lamp versions are very durable but extremely tough on lamps, I haven't pulled out a ton of lamps older that 2006. I have found a couple from 2003 and even one from 1999, but these ballasts last a long time at the expense of the lamps.

Sadly the lamp quality from Philips, Sylvania, & GE have been very poor since they were required to reduce mercury levels even more since about.. I believe 2019/2020 or so.  While I rarely see the newer Sylvania lamps go mercury starved, I am seeing them and GE and Philips go EoL much sooner.  Its a real shame.  You can only reduce the amount of mercury so much in a lamp before it no longer will work right. 

likely a "dumb" question where does the mercury go when they become Mercury starved?
 

There is a single digit of milligrams of mercury in a tube with relatively large inner surface. Mercury can be absorbed in phosphor layer or on a glass surface, or can be chemically immobilized by reacting with phosphor, reactive gases remaining in a tube, or again, glass surface. Some quite obscure things can happen, for example if a tube has been incompletely bombarded and outgassed, some residual water may remain inside, water will slowly hydrolyze glass surface, liberating sodium ions. Sodium will react with mercury forming intensely black sodium amalgam, which will stain the glass and consume available mercury.

thanks for the reply
I have heard of running tube on DC power will draw the Mercury to one end but never heard where Mercury went in a starved tube