Automatic fridge and freezer temperature log for food safety

Every two hours during opening hours, keep the cold storage temperature log for my food business so nobody has to fill in a clipboard by hand. Use Home Assistant List States to read the current temperature for every refrigeration unit I care about: walk-in fridges, freezers, cold rooms, prep fridges, and display cases. I will give you the list of sensor entities, a friendly name for each unit, and the safe temperature range it must stay inside. Treat that list as the definition of what gets logged, and default the schedule to every two hours between 6am and 10pm.

On every run, for every unit on that list, append one row to my Google Sheets compliance log using Append Values, even when everything is fine. The continuity is the whole point, because an auditor wants an unbroken record rather than just the exceptions. Each row should carry the timestamp, the unit name, the sensor it came from, the reading with its units, the safe range for that unit, and a plain PASS or FAIL marker. Add a short note column for context, for example within range, above range with a defrost suspected, or sustained excursion already alerted.

Before you raise any alarm about a unit that reads out of range, call Home Assistant Get History for that sensor over the last four to six hours and look at the shape of the curve. A freezer that jumps up for twenty minutes and slides back down is running a normal defrost cycle. A fridge that rises while the back door is open for a delivery and recovers once it closes is not a failure either. Threshold-only alerts cry wolf constantly and staff learn to ignore them, so your job is to be the judgment layer that separates a brief self-correcting excursion from a genuine one.

Treat it as a genuine excursion only when the unit has stayed outside its safe range for a sustained period, defaulting to at least 45 minutes, and the history shows no sign of recovery. When that happens, post to Slack with Send a Message in my food safety channel, naming the unit, the current reading and how far outside the safe range it is, how long it has been drifting, and whether the trend is still climbing or has levelled off. Write it so a duty manager can act on it in one read.

Do not repeat the same alert on every subsequent run. Before posting, use Google Sheets Get Values to read the recent rows of the log and check whether you have already alerted on this unit for an excursion that is still ongoing. If you have, log the row as normal and stay quiet. Speak again only when the unit recovers back into range, in which case post a short all-clear, or when the situation gets materially worse, for example it crosses a further threshold or has now been out of range for several hours. Always mark in the log row when you alerted, so the next run can see what has already been reported.

Stay read only. Monitor, log, and alert, and never attempt to change a device setting or tell me you have adjusted anything, because this workflow cannot control physical equipment. If a sensor is missing, unavailable, or reporting a stale value, treat that as its own problem: log it as a gap in the record and mention it in Slack, because an unavailable probe on an audit day matters as much as a warm fridge.