My IoTaWatt unit has been plugged in and running for the best part of 18 months (Thanks @overeasy ) with barely a ripple.
As part of that it uploads to both emonCMS and PVoutput.
Recently there has been MANY lines in the log like the following:
6/15/21 00:49:07 PVoutput: HTTP completion, upload -4
6/15/21 00:49:30 PVoutput: HTTP completion, upload -4
6/15/21 00:49:53 PVoutput: HTTP completion, upload -4
6/15/21 00:50:16 PVoutput: HTTP completion, upload -4
6/15/21 00:50:39 PVoutput: HTTP completion, upload -4
6/15/21 00:51:02 PVoutput: HTTP completion, upload -4
6/15/21 00:51:25 PVoutput: HTTP completion, upload -4
6/15/21 00:51:48 PVoutput: HTTP completion, upload -4
6/15/21 00:52:11 PVoutput: HTTP completion, upload -4
6/15/21 01:00:08 PVoutput: HTTP completion, upload -11
6/15/21 01:10:07 PVoutput: HTTP completion, upload -11
6/15/21 01:30:07 PVoutput: HTTP completion, upload -11
6/15/21 01:40:08 PVoutput: HTTP completion, upload -11
6/15/21 01:50:07 PVoutput: HTTP completion, upload -11
6/15/21 10:00:08 PVoutput: HTTP completion, upload -11
6/15/21 12:00:09 PVoutput: HTTP completion, upload -11
6/15/21 12:30:09 PVoutput: HTTP completion, upload -11
6/15/21 12:40:10 PVoutput: HTTP completion, upload -11
6/15/21 13:00:09 PVoutput: HTTP completion, upload -11
The data still makes its way to PVoutput and there is no obvious impact on performance but just a LOT of extra lines in the log.
I have not made any config changes since I set it all up many months ago and I’m not keen on messing with it unless it’s actually broken.
Any clues on how to trim them down a bit?
I’ve been seeing the same thing with my units. The code -11 is a timeout. As you say, it eventually retries and succeeds. I notice that with your system and mine, these events usually happen on the hour. I believe it’s something happening with PVoutput. They have a flow control where users who exceed the hourly limit are suspended until the start of the next hour. While that may reduce transaction during the hour, it seems to me that it has the potential to unleash a flood of transactions on the hour. That may have something to do with this.
For now, just live with it. I’ll contact PVoutput to inquire about it, and also look into suppressing the bulk of these recovered error messages.