My temperature monitoring project has been running very well, especially since I corrected the humidity sensor readings by applying a correction factor. And the upstairs temperatures did drop below freezing a couple of days ago, so the tasks of shutting off the pump, draining the plumbing, and adding antifreeze in the traps were absolutely worthwhile.
But I decided that it would be nice to know when some threshold temperatures were reached without checking the web page. I added code to the python3 logger script to write a file when a cold or heat threshold was reached. It also sends an alert email message when this happens. The web page that displays the temperature and humidity graphs looks for this file and indicates that a threshold has been crossed for cold or hot temperatures.
It seems the DHT22 sensors can waver back and forth a few tenths of a degree, and if that happens right at one of your threshold values, you will get an annoying collection of emails. To address this, I implemented a buffer mechanism. Now, the heat/cold alert will only happen when the temperature crosses the threshold and stays there for 3 consecutive checks. Since I’m logging the values every 5 minutes, that means the temperature must remain either above (heat) or below (cold) for at least 10 minutes before an alert is sent. When the temperature rises above the cold threshold, it must remain above the threshold for 3 consecutive checks before the alert is rescinded. Rescinding a heat threshold works the same way.
I also added some code to rescind any existing cold or heat alerts when the logging daemon is restarted. Since you can edit the threshold temperatures, it seems reasonable to start clean. If an alert should be issued, it will take just a few minutes to get past the buffer mechanism.
So far this is working well. On my next trip up there, I’ll recheck the sensors using the hygrometer and see if any alterations to the correction factors are needed. If they are, it’s a trivial change to make.