I agree if I’m the only one having these problems, this may be locally related. However, out of my 337 things in total, the Tibber live items are the only ‘troublemakers’ right now. Please be ensured that I am very very thankful for your effort and everyone else making OH to what it is today and all the functionalities that it provides.
I’ll be on vacation next week and need the Pulse data to be rock solid feeding my rules during the absence. Therefore I switched to IO Broker and MQTT for acquiring the data. Happy to investigate further after my vacation.
EDIT: in the future it may also help to have some more debug-information to solve situations like this.
Actual status of my configuration:
As far as I can see, everything went well with aquiring data via IO Broker / MQTT for the week of my absence. No interruptions.
I stopped my OH installation last Sunday, deleted tmp / cache directory, switched to the Tibber API binding, deactivated the IO Broker instance and started OH again.
After one week, the results are 5 outages where all outages could have been recovered by automatically restarting the Tibber API thing.
Until now, the “thing-restart-workaround” is still needed for my configuration.
Not directly connected with this binding but maybe of interest for other users:
As the Tibber Pulse is using the optical interface to my energy meter, I had to migrate away from the current DSMR solution to a local Pulse-querying solution.
After setting the pulse up for the webserver to be always enabled, not just for setup, it is possible to query the data the pulse sends to tibber. This python script polls this data, transforms it and publishes it to a local MQTT broker.
2024-05-14 07:01:00.682 [WARN ] [core.thing.internal.ThingManagerImpl] - Initializing handler for thing 'tibber:tibberapi:Tibber' takes more than 5000ms.
since switching to OH4.2 I’ll get timeseries log entries, even though the value of the “current price”-item did not change. Every minute that is. It’s filling up my log, so is it possible to change that in the binding to only log timeseries information, if the value received changed?
see the parallel thread here:
Also using the Tibber binding and got these log messages every minute. I agree it should only update the (forecast) item in case of a changed value. You can reduce the logging however by setting the update interval in the item configuration to a higher value so it will only log at the requested interval. Energy prices get update once a day anyway.
I am very happy actually that the binding supports forecasting now so I can display future values in graphs and make calculations based upon look-ahead energy prices ! Works great with in memory persistence.
i just filed a feature request/improvement on github:
to add more context:
this was the origin, that I got too many events.log entries:
and then I changed the “refresh interval” to 25 minutes, which led to the behaviour, that all non-live channels refreshed “randomly”, but not “on the hour”:
So my suggestion is to change all the hour-related channels “on the hour” as described in the github-issue.