The storage data indicates that the system is fairly large

Upon reviewing the new Storage Insight feature, I noticed that the system takes up more than 40 GB of space.

That’s too much space, in my opinion. How can I see what’s taking up the space?

My installation:

  • HAOS (ova) on Proxmox 8.4.5.
  • Core 2025.9.1
  • Supervisor 2025.09.0
  • Operating System 16.2

My NUC with HAOS on bare metal has 18.9 GB, so it must be counting other things to the system. I do hve my database in a Mariadb addon, so the recorder database might explain the difference. Although I must admit I would not expect that to be counted as system.

If you want more detail see my guide on how to check used storage.
You can see the paths it checks here.

I was running out of storage space until I moved Frigate to its own server.
Now:

I found the following:

/mnt/data/supervisor/tmp 24.6 GiB
/mnt/data/docker/overlay2 26.8 GiB

In tmp, old broken backups are being stored, so I deleted them by connecting via ssh using this tutorial:

cd /mnt/data/supervisor/tmp

ls -la

rm -rf tmp*

As a note to others investigating a large backup (not bragging about size but 18.9GB here!).

I found .git/objects/pack was 18GB. An automation pushing github backups of “/config” was growing over years of tinkering, storing every commit ever made. Additional issues with .gitignore and some clumsy git usage adds up. Oops.

In my case, old branch movements and dangling commits were neither intentional nor valuable.

git reflog expire --expire=now --all
git gc --prune=now --aggressive

And afterwards:

cd /homeassistant
du -sh .git

24.6M	.git

Hope this helps others frustrated by backup storage.

how did you get SSD lifetime in your setup?

This really needs to be addressed with built in functions rather than getting hacky…

EDIT: And on another system…

another thing that took up a lot of space in my install were ESPhome build files, some of those are required, but there were lot of unused things so using that CLEAN ALL FILES button in ESPhome helped a lot (it will redownload only required things next time you compile update for your devices)

another thing were old docker images docker image list
to clean them up you can use docker image prune… it only removes images of apps that are no longer used, but leaves old versions of used apps
to remove old versions of used apps run docker image prune --all (it removes all unused images, containers, networks)

Hello,

I have the same problem. ESPhome CLEAN ALL FILES saved arround 2%
But where can I enter the docker image list or docker image prune command?
I tried in terminal but it says
bash: docker: commmand not found

Best regards,
Stefan

Use the Advanced SSH App with disabled protection mode. See my link.

Yes you can see at my first site I have 0GB on my App Data, that’s after clearing ~5.5GB from ESPHome! Second site I hadn’t cleared that yet.

As for docker containers, looks like you’re right and I count about 8.6GB here!

(For some reason I cannot copy text out of Advanced SSH, so a picture will have to do)

By my maths, that’s ~8.6GB worth… which matches what is showing on my first server for “System”.

I’m not sure if there’s anything to prune here though? I currently have Music Assistant disabled, along with Samba disabled, but I don’t want those deleted accidentally in a prune, so not sure if it’s safe to try.

If there is a bunch of unused old container images, there definitely needs to be some focus on HA cleaning up after itself.

It’s not safe at all. Any addon which isn’t currently running will be removed if you run the prune command.

Ask me how I know :man_facepalming:

Thanks, this helped. I cleared the ESPHome build environment and that gave me back 15GB of space.