Why would HA try to store the history data that way? I thought it obvious that this would get performance issues on big setups.
Knowing this now, I see why I have those performance issues with displaying graphs of more than say half a day.
Doing a CLI based file search to query data cant be producing performance records.
I would never have imagined that HA does use such a db design… I really can say im totally flabbergasted.
Thanks for the Influx hint, I was already considering MariaDB.
So what is the user case difference between Influx and Maria here on HA?
I know influx is a special design for time to value data and as such the perfect match for our HA data while Maria is just the usual SQL.
The thing is, I have a Maria already up and running for other stuff including snapshots and so on, I would just need to add a homa-ssistant-data db and the user for HA.
So what is the difference between the HA data going to Influx and Maria?
I found this, but its not clear to me what the performance issues would be.
Performance wise and DB size wise.
THX