Add rolling log for crash report

We saw a mysterious crash today, where 3 of the V1 sensors went down within 5 minutes of each other, when they were also started up within 5 minutes of each other, the night before.

They rebooted normally and the cause is not apparent.  We should save the last X lines of logging in a rolling window of space so that they can be transmitted to the api endpoint on reboot.