Error persists on entire layer where a new retraction pattern is used, therefore it is very unlikely to be filament humidity or a gcode issue, but filament settings.

What is your retraction length and speed? Do a cold pull to rule out debris. Reduce your retraction until stringing occures, observe if the error repeats. Temperature could be playing a role, but if your layer adhesion is up to expectations, I wouldn't change. Also observe your nozzle and make sure it's taught, I've been surprised how fast some of my brass nozzles have worn out even with non-abrasives, and causes a similar symptom of randomhish failures, especially if loose.

PS: sometimes a couple cold pulls are needed and try to hit different temps when you pull. I've noticed since PETG doesnt stick to PLA, to use PLA as a cold pull you actually need it warm enough that the PETG bits will flake off the inside