Hi, I am experiencing more “Failed to set inverter” errors in my activity log since enabling Agile. I suspect that this is partly because more instructions are being sent and therefore a higher chance of something going wrong.

Does WW have any mechanism built in to retry sending the instruction if it fails? It would be good if it could capture the error, wait 60 seconds and retry and loop around that process until it gets a success message, I would think you would probably want to limit the retries to 3 and then give up. This would give a a few minutes for whatever the issue is to sort itself out (be that the api, my dodgy internet connection etc).

Alternatively, Growatt (not sure about giv) allows you to set multiple schedules so you could send the next schedule in Slot 1 and the subsequent schedule in Slot 2. If the subsequent instruction fails to send it will still be actioned as it is already set in slot 2, if it is successful, what was in slot 2 would be sent to slot 1 and slot 2 populated with the next subsequent schedule. (Not sure if that makes sense…)

Anyway, I appreciate this is all in beta and your hard work on the continuing development, so know there will be occasional blips. I suspect if WW goes into full production users might be less forgiving if they miss out on cheaper electricity rates due to the inverter failing to be set, so worth considering.

    jjb We do have a retry mechanism set up - 3 retries and a backoff factor as well.

    There are different reasons for this error, sometimes, especially on the Growatt side, users have other timeslots configured which have overlaps with the ones we send/set. That's usually the most common problem.

    As for the Inverter Timeouts, again, retries are in place, and it's also worth noting that some people have a lot of these, whilst others see none, which could also be related to Internet/Wifi connectivity to the inverter dongle.

    For some extra context, here is the log for the top number of errors per user. Each line is a single user.

    Some of those highest ones are likely not very active on WW, and have conflicting schedules. So all in all, we're talking a small minority that have issues.

    • jjb replied to this.

        admin Good to hear you have retries already in place.

        Last night I got two more of these errors which meant my battery did not charge at all. Are you able to see more detail about why it is failing? I would like to check if there is anything this end I can try and change to get it behaving more consistently?

            jjb Thanks for raising this.

            We do see this error when there are conflicting timeslots already set up at the Growatt end. And also down to general intermittent behaviour of the Growatt API/backend.

            We had a better look at our retry policies, and due to the HTTP status code coming back as a 200 (Success) - the retry policy didn't actually kick in for Growatt schedules when encountering this error!

            This should be fixed before midday.

            • jjb replied to this.

                Just for information: Growatt have an annoying habit of setting "Grid first", "Time slot 5" as "Enable". This causes a conflict when trying to set "Battery first" parameters. It is a bug, but I dont know if it is in the Shine app or in the inverter firmware.
                When I change anything, I disable slots in "Grid first" and update (Yes) before changing "Battery first". If I forget, I get an error.

                Thanks for the update, and the tip on the Grid First issue. I will keep an eye on it and let you know how it goes.

                7 days later

                admin Just wanted to let you know that since you made this change instructions have been 100% successful. Thanks for the fix.