Presently the logging of vacuum_monitor is set up to not reload the .csv for each log event. This is normally optimal for speed, but it leads to the log becoming disrupted when new dictionaries with different instruments are appended to an old log. Should add logic which auto-detects if instruments have changed and reloads.
Temporarily fixed problem by adding a flag in vacuum_sentinel to reload dataframes during testing periods. Checked that this does not give a perceptible speed hit, even with months worth of data.
Presently the logging of vacuum_monitor is set up to not reload the .csv for each log event. This is normally optimal for speed, but it leads to the log becoming disrupted when new dictionaries with different instruments are appended to an old log. Should add logic which auto-detects if instruments have changed and reloads.