Clearly document impact of (irreversible) admin actions

[Nothing4You]

Certain admin actions like purging objects or banning users with content removal have destructive elements.
Their impact should be clearly documented.
For example, instance-banning a user with content removal will erase all of their uploaded media, generated post thumbnails, as well as their bio.
For purges, it should also be documented which data remains, as the username remains occupied.

It would probably be a good idea to start with a general documentation of all admin actions and documenting the expectable outcome.
It might be a bit challenging to keep this up to date though, as it's easy to forget updating documentation here after the code in the Lemmy repo changes.

Maybe this could also be documented in the API docs directly for the detailed descriptions and represented in a more accessible way through the docs, as admin documentation shouldn't require digging through API docs and identifying how endpoints relate to actions within Lemmy-UI. API docs will include the URLs, parameters and responses, but admin documentation would focus on the actions available through Lemmy-UI and describe how they're used in that context.

This could also eventually be extended to mod actions and even user actions, but I think admin actions are more important due to the potential destruction of data.

[dessalines]

I don't know that this would be that useful in the docs, bc ppl aren't going to read through the docs before using lemmy.

It'd be most useful to go through these in lemmy-ui (and I think most of them already have permanent removal warnings.).

The only ones I can think of are:

• All purges
• Ban + remove data
• Deleting account
• Removing images

[Nothing4You]

I agree, it would probably make more sense to document them directly in the API docs and in Lemmy-UI, which hopefully other clients would follow in a similar fashion to include documentation.

For Lemmy-UI it might also make sense to have some kind of help texts with a (?) or something like that near actions that aren't fully self-explained, even if they are not necessarily as destructive as the described ones.