I think before releasing the animal back into the wild and especially while it is still apart; with power on and some kind of signal going (because the fault you report is signal cutout and not a great big static crunch--meaning you NEED the signal to detect the fault) I would try to flex the board downwards and upwards maybe 1/8"...right in the area of that connector. See if you can get cutout. Give it just a little more force than it should ever see but nothing violent.
PC boards can generally withstand lots more flex than you might imagine, but all it takes is one trace, maybe a little thin on the plating, that somebody accidentally scored with the blade of a screwdriver during handling.
I once worked in a plant where they made OEM CRT displays...for ATMs, and those little 5" displays you used to see on cash registers. One particular model, by far the most complicated they made, used a two-sided PC board. All the others were one sided. As it turned out, the artwork between the two sides was slightly misaligned. (supplied by an outside supplier) The unit used "stand-up" type pots for V Hold, Brightness, etc; etc; and the legs for the pots ended up creating all manner of shorts to the ground plane on the side of the board where the pot connections were NOT made. These things would fail off the assy line, go back for repair 3-5-8 times. I ultimate proved that I could fix all of them by unsoldering the pots and lifting them off the board 1/32". It took 5 months of the engineers (I was just a lowly tech) laughing at me to show them what was what. The head engineer then sat me down and offered me a 25 cent per hour raise. I replied, "OK, I am going to leave work today and every day hereafter at 2 pm until I find a real job". He practically fell over. 'Course, that was LA in the eighties, and if you could show up two days in a row someplace and not smell bad, they would hire you. Nothing like today.