Looking at those traces the PV does not have any spikes, it follows what the CV is trying to do so apart from those very short spikes looks like the PID settings are wrong.
I would personally say that there is unlikely to be a wiring problem or thermocouple. obviously, traces are fine but the update time might be a bit slow to catch very large but lengthy ones, in saying that, it is unlikely it would not see at least a few. apart from those unusual spikes it looks like the classic symptom of poor tuning to the process, although these heater cells may be the same does not mean that they respond the same, after all the PID is tuned to the reaction of the feedback an example is two heaters of the same type, one might take longer to reach temperature than the other (perhaps due to mechanical factors like blocked ports or a partly shorted electrical heating element (had that one before), in that case the PID settings would possibly need to be changed to overcome temporarily the problem.
Indeed, in the late 90's I had some Eurotherm controllers on a Rossi catelli core direct steam injection system to cook sauce, this was cooked at 101.0 degrees C (no it does not boil it is under pressure) the process was critical +-0.5 Deg C according to the process development guys, the problem was that over the week the injection core would slowly block, the controller then struggled to control at the required max temp deviation allowed, we put the controller into Adaptive tune so this overcome the problem, the engineers then produced a PPM schedule at the weekends to strip the injector & clean it. AS for those spikes on the CV no idea. Me personally, I would if possible swap the offending control loop with another (assuming they are identical) & see if it happens on the existing PID loop or the other, that way it may give an indication of if it was wiring/heater bank or perhaps noise within the cabinet/PLC.