Sensor #35 was built on Oct 2, 2018, and first went into service on Mar 9, 2019 until two days ago.
This particular sensor has had 3,195 power cycles, for a total powered time of 1,954,623 milliSeconds, or 32 minutes and 34 seconds spending the entire time in one pot.
You can see the corrosion on the left probe.
A closer view...
The data shows it had no effect on the sensors performance in that short time.
I will continue to monitor the condition of sensors as they are removed from pots to determine at what point it begins to affect the readings. It will be simple enough to automatically determine the point just before failure, and prompt me to replace the sensor (or at least the probes)
This leads to a design change for any future sensors...
Right now I plug the ends of the terminal block, where the steel rods are connected, with hot glue to ensure it is waterproof. Future versions will have heat shrink applied first, then hot glue, then more heat shrink. This will keep the hot glue out of the screws and threads, allowing me to replace individual probes.