I use Jetson Nano for a data collection and GPU inference task. These tasks are perform in parallel, so while the Nano captures de data (images) I also do inferences with the images (yolov5 in pure torch1.81), save data to a db and delete the infered image.
This pipeline has gone perfect for some days, but after 10 days maybe it started to fail, I check with last reboot if the nano reboots and it seems that it randomly does it while doing the job (random because one day it reboots after 5 hours of work, and other it reboots after 30 minutes).
I am using the power supply requested (5v-4A), also the fan is setted to default. the environment where the jetson is working is a laboratory (monday to friday 9am-6pm with air condition, others no and closed with no windows)
Maybe this is due to overheating? Is there a way to check a log? like heat log or error before rebooting?
Thanks in advance.
PD: Some Jtop Captures: