I used authentik first as it was suggested by a friend with whom we are sharing the vps, and so I kept using that also for another personal vps.
It is a good combination of robustness and flexibility for me, plus we needed a user dashboard or at least a list of apps. Custom workflows, blueprints and stuff were a bit of a pain to understand, so I just copy them between the two install I have because I keep forgetting how everything works. It's also quite resource-heavy, so I might try something different in the future.
Not really automation, but I will copy-paste a comment I made in an older thread, maybe it can still be useful :)
I download books through Z-Library, I tried in the past lazylibrarian and readarr/bookshelf for automation, but I always ended up spending a lot of time on refining the search for non-english books or specific editions, which was just way easier in Z-Library.
I tag/update metadata of books using Calibre of my laptop. I prefer to manually do that because while it took a few days to do it for my whole library, I don't really add books in such a high number that I can justify automating it or having a Calibre container on my homeserver. (I know Calibre Web Automated will do this, but see below). I use Hardcover to track my readings, so Calibre with the hardcover plugin makes it easy to find ids or HC editions for integrations with other apps I use.
After moving between Kavita, CWA and Booklore, I set up on Komga for library management. I found Komga's koreader sync way more stable and mature than the other options. I am a bit annoyed by the fact that (as in Kavita) each book/series folder is a series, but at the end of the day the pros outweighs the cons for me. I can manage custom reading lists or series, so it's not really an issue personally.
I would go for hedgedoc if the team can switch to markdown, otherwise Cryptpad to stay in an office-like suite.
If you already have nextcloud then that + collabora should be good enough. Otherwise installing the whole nextcloud suite just for collaborative editing seems like a lot imho.
Komga is always mentioned as a good option for this. I use it for books, manga, and comics, but my requirements are quite minimal (koreader sync, opds, and mihon), so it might not work for you exactly.
I do metadata fetching using komf, documentation is a bit lacking but once setup is pretty good (I use it for both manga and comics) and it does everything automatically. There is also a browser extension that provides a UI for it. For comics sometimes I use ComicTagger for manually reviewing tags too.
I would just read carefully the folder structure Komga expects if you go with that, to avoid having weird surprises :)
Going a bit against the common suggestions here but my "workflow" is like this:
I download books through Z-Library, I tried in the past lazylibrarian and readarr/bookshelf, but I always ended up spending a lot of time on refining the search for non-english books, which was just way easier in Z-Library.
I tag/update metadata of books using Calibre of my laptop. I prefer to manually do that because while it took a few days to do it for my whole library, I don't really add books in such a high number that I can justify automating it or having a Calibre container on my homeserver. (I know Calibre Web Automated will do this, but see below). I use Hardcover to track my readings, so Calibre with the hardcover plugin makes it easy to find ids or HC editions for integrations with other apps I use.
After moving between Kavita, CWA and Booklore, I set up on Komga for library management. I found Komga's koreader sync way more stable and mature than the other options. I am a bit annoyed by the fact that (as in Kavita) each book/series folder is a series, but at the end of the day the pros outweighs the cons for me. I can manage custom reading lists or series, so it's not really an issue personally.
Also Komga works pretty well with Koreader. I moved my books from Kavita to Komga and I found it to be more reliable regarding the two-way syncing progress (at least for me).
I generally don't update automatically, I currently use WUD. It works fine for image checking and notifications and had no need to change it for now, but I am thinking of trying dockhand too.
Someone recently recommended me Textadept: https://orbitalquark.github.io/textadept/. Haven't tried it on linux, and I am not really using it, but the interface is clean, it also has a CLI, and I thought I could give you another option :)
Thanks for the thorough reply! I didn't know about Inav, but it looks very interesting. I agree on the Grafana stack, it's not something I really need now, and if I have to inspect single containers I can go for something like Dozzle.
About crowdsec free plan, looking at the pricing page, I see that the community plan has unlimited remediation components and 3 blocklist + unlimited scenarios, or am I looking in the wrong place? (honestly that page is pretty confusing)
Thanks for the input, yes I was mostly thinking about hedgedoc, that doesn't have parsers or anything. I need to delve more into crowdsec logic and rules before trying to do my own thing, for sure. Thanks a lot tough, I followed your advice and I got Crowdsec working on both Authentik and Forgejo :)
I used authentik first as it was suggested by a friend with whom we are sharing the vps, and so I kept using that also for another personal vps.
It is a good combination of robustness and flexibility for me, plus we needed a user dashboard or at least a list of apps. Custom workflows, blueprints and stuff were a bit of a pain to understand, so I just copy them between the two install I have because I keep forgetting how everything works. It's also quite resource-heavy, so I might try something different in the future.