Hello @OddJob,
I it’s been awhile. Like they say, no news s good news. Looking at that log is a trip down memory lane.
Looks as if your IoTaWatt run straight through March and did a routine six week restart.
Then the wheels came off the wagon when you started using influx. I will admit that my primary home unit does all of EmonCMS, PVoutput and influx, and I do occasionally get a heap degraded restart. But maybe once a month. Every 2-12 hours is a problem.
First, I don’t think it has anything to do with the vintage of the unit. It seems to have been brought on by influx.
Looking at your status display after 6 hours the heap is still healthy. So it must crash because of some stimulus. Because I’ve had a bout with this issue, there is an undocumented diagnostic feature that can be used to add an available heap measurement to your influxDB measurements. To use it, you add the Json object “heap”:true to the influx configuration.
This involves editing your config.txt file with the file manager app. If you make a mistake and cause your config.txt to be invalid Json, your IoTaWatt won’t boot and you will need to remove the Sd card to fix it on another computer. I always use a Json lint program to check a config after I edit it.
Here is an excerpt from one of my config.txt that is monitoring heap:
“influxdb”: {
“type”: “influxdb”,
“revision”: 50,
“heap”: true,
“postInterval”: 10,
“bulksend”: 1,
If you don’t want to edit the config.txt file, you can post it in a PM and I’ll do it.
Once it’s been setup, there will be a measurement called “heap” with your tagset and having a field “value” equal to the average available heap during the period since the last heap measurement.
Doing this and then plotting the heap measurement might shed some light on this issue. The restarts will be apparent (at least to me) and the heap usually degrades in noticeable steps.