Today I was smoking a turkey so I figured it was a good time to do a live test of my project. There was good news and bad news. I think it's illustrated quite well by the following graph:

On the good news front:

• While I didn't do a minute-by-minute comparison (because I was busy making some side dishes), the temperatures it measured seemed to be fairly accurate and reasonable.
• The temperature was able to travel via MQTT to the InfluxDB and Grafana Docker containers.

On the bad news front:

• I kind of already knew this from experience with my cell phone, but WiFi reception is garbage on the BBQ patio. The 45° line from 15:45ish to 16:28ish is because the device wasn't reporting back. I went out with my laptop to see what was happening and it wasn't able to connect to the wifi. I lifted it up in the air (connected to my laptop) and it was able to connect. Then I rigged it to an extension cord in the tree (rather than near the ground as in the opening picture) and I guess it was able to do a bunch of data connections. But I wasn't done smoking at 1700, yet that's when I have the final data point.

So it seems like I've at least got a WiFi problem to solve. But when I've left it going in the computer room in the past I've seemed to have seen it conk out sooner than I'd prefer for a smoker application. (Conk out as in refuse to continue sending data or losing the connection to the SSID, not turn off or something.

I've added in a timestamp to the MQTT to influx Python script so that I can truly see when the final message came through. While testing at my desk I may also need to temporarily lift the restriction I made for not sending bogus temps in order to know for sure whether it's lost the WiFi or whether it's just not finding legit values to send.

So I'm pretty encouraged, even if things aren't right where I want them to be yet.