Help on History - Recorder

1

Team,
sorry to bother but I need help.
I’ve enabled MariaDB for history and recording and is working fine, I do see the DB growing :slight_smile:

I’d like to track everything even unavailable etc…
but when I try to see some devices I do not get any info.

here is my config:

recorder:
  db_url: !secret mariadb_url
  purge_keep_days: 730
  auto_repack: true

I’ve removed any exclude: or include: and also I’ve removed the history:
my expectation is to have everything saved

but:

image
image1377×516 19.4 KB

this device has gone online and offline multiple times in the last day so I would have expected to have “unavailable” and available or things like that.

What am I missing?

Please note 730 is ok :slight_smile: I want to track 2 years :slight_smile:
It’s on an external MariaDB with enough processor and storage :slight_smile: I’ll review the number in case the size becomes tooo big

2

Recorder - Home Assistant

It records sensor values, not availability or binary or toggle. How can the database possibly know something is not available, unless you build infrastructure that polls everything and puts it on a heartbeat or something. Items check in when they check in.

Storing 2 years of data will make that thing slow to respond, laggy, and hugely prone to DB corruption because it takes so long to do things.

It’s not the speed of the equipment running the database, it’s the HA server that has to sort thru the data.

Recorder - Home Assistant

The data after a short time is aggregated into like 5 minute or something blocks so it’s not real data anyway.

History - Home Assistant

Please read the HA Docs on what the database is and is for. Short term handling of the current state in order to make decisions on current actions.

If you want to store everything have a separate external time based database like influxdb and HA will work as intended in its command and control role.

Personally I just delete the database when it gets too big because knowing what the temp of the furnace was 2 years ago last Tuesday has no effect on my life.

1 Like
3

It records sensor values, not availability or binary or toggle. How can the database possibly know something is not available, unless you build infrastructure that polls everything and puts it on a heartbeat or something. Items check in when they check in.

Status and availability is an information that could be recorded, is there a way?
In HA I can see when a devices goes unavailable and then back.
In activity I can see:
**

image
image582×246 6.91 KB

**

It’s not the speed of the equipment running the database, it’s the HA server that has to sort thru the data.

Isn’t this done by the DB? HA should query the data it wants and get it back, am I wrong?

I’ve read the Recorder and History sections but are not clear to me, or at least maybe I’ve not understood them

If you want to store everything have a separate external time based database like influxdb and HA will work as intended in its command and control role.

not clear how to that… can you help me more?

Personally I just delete the database when it gets too big because knowing what the temp of the furnace was 2 years ago last Tuesday has no effect on my life.

hahah true! but if you want to use AI together with external info to define patters and optimize costs, for example, that is needed.
Once the db is not an issue as is managed on a high perf entity I can track everything and use what I need, otherwise I would have to use the include exclude to be more precise

thank for the help
M