OH5: Influx 1, 2 or 3?

Setup, Configuration and Use Beginners
1

I have a Pi 4, 8GB RAM and installed OH5.0.1. I try now to install Influx and see that there are two links for Influx 1 and 2. When I follow the link I recognize the V3 Core is as well out. I am confused which version is the best to use, especially since the documentation mentions only V1 and V2.

If I install a new database should I stick to version 2 or use there new V3 already? If so, are the plugins (like Grafana) for Openhabian working with V3? Or should I use the provided ones from Influx?

The other concern is that I am not clear if my hw is suitable for V3. I did not find any hints on that topic neither.

2

The openHAB InfluxDB Addon is not capable of using V3 yet. See

Afaik openHABian will install V2.
Grafana is fine with each version, you only have to configure it correct :slight_smile:

3

According to the documentation it says

First of all, you have to setup and run an InfluxDB 1.X or 2.X server.

Which would mean, it won’t install the database, only a connection interface. But perhaps I’m wrong and reality looks different.

4

I was migrating to openHAB 5 not so long ago and openHABian installed V1 (1.12).

5

V3 Core is not suitable for openHAB. It has a 72 hour limit for queries.
There is V3 Enterprise free at-home license without this query limit. They say this home license is rate limited. Whatever this means.

My opinion: V1 is still ok for openHAB. In future you have to look for another product.

6

Do you accidentally know if there‘s a good migration instruction somewhere for OH?

7

I’m still running influx v1 on my OH installation. Mostly because it is capable of using retention policies and continuous queries, which automatically downsample data to become a rrd-style database (higher granularity for recent data, lower one for older data). This automatically works for all measurements and also for newly added measurements without user interaction.

I’ve looked at some other time-series databases (timescaleDB, influx v2, even postgresql), and none of them support this. Yes, you can downsample data, but it needs to be done for each and every measurement, which kind of defeats the purpose.

Now influx v3 apparently has that feature with the downsampling plugin, so I’m eager to see if and when OH will support it.