Calculated Field Missing for Burt158_boiler_gas about bmon HOT 2 CLOSED

It is likely that this occurs during periods when the data collector cannot reach the BMON Internet site and the readings are queued up locally. When the readings do get posted in bulk, the calculated readings only go back a certain amount of time. So, if Internet outage was longer than this, the calculated readings will not fill in the gap.

I determined this by looking at the Burt158_firing_rate data and the Bur158_boiler_gas data. See the IPython notebook C:\Users\Alan\Dropbox\AHFC\BMS-3\data_analysis\calc_field_problems.ipynb. I noticed that during periods when the calculated field was missing, the firing_rate field was not evenly spaced in time, indicating that a reboot was occurring periodically. The reboot was occurring because it couldn't reach the Internet. This was clearly verified by looking at the cron_15m log file: http://www.analysisnorth.com/marie/files/Burt158_2014-11-12-0646_pi_cron.log

One solution to this is to have the calculated field process go back further in time so that is likely to fill a large gap of missed readings. Fix that in this file:
https://github.com/alanmitchell/bmon/blob/master/bmsapp/scripts/calc_readings.py
by changing the 3rd parameter to the calcreadings.CalculateReadings() call.

from bmon.

This also indicates that Burt158, at a minimum, is experiencing long periods without Internet connectivity. The cron_15m log files shows the Pi trying to restart the cellular network and then rebooting the computer about once per hour, to no avail for a number of days. Should try to figure out whether this signal strength or some other issue. And, reboots do not seem to solve the issue.

from bmon.