Looked at the last 6 locklosses from Observing.
Could see nothing much in all but one of them, 1443315128 where ETMX_L3 and DARM get noisy in the 70ms before LL, plot attached.