I’ve always been conservative about what kind of services I host because it takes time to get them set up. For example, there’s no reason for me to set up music streaming when I only ever listen to music on my phone and all my music files are already on my phone. On the other hand, it’s a good learning opportunity to set stuff up and have to fix it when it breaks. What do you think?

  • sneakyninjapants@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    4
    ·
    2 years ago

    Oh I have quite a few that I’ve set up then pulled out of service for various reasons. I’m always evaluating potential use-cases for new services and if a different service would better suit my needs than what I have deployed currently. It’s definitely a hobby.

    Some container-based projects that I’m loosely tracking updates for and have deployed, but since, have pulled out of service (non-exhaustive):

    Media:

    • Calibre / Calibre-Web

      • Supplanted by Audiobookshelf. For tagging and book conversion I just temporarily install Calibre on my Workstation when the need arises.
    • FreshRSS

      • Did not end up using it for much at the time. Re-evaluating if I’d like to stand it up again.
    • Plex

      • Plex is nice but too many drawbacks that don’t work for me. Supplanted with Jellyfin.
    • Overseerr

      • Didn’t have much use for this, but it will likely change soon. Since I’ve stuck with Jellyfin I’ll be going with Jellyseer if I decide to stand up this kind of service again.
    • Libreddit

      • Rarely made use of it. Nice project, but it’s not feasible now for obvious reasons.
    • Miniflux

      • Same as FreshRSS, though I’m a big fan of go and rust projects in general so this is the one I’m more keen on re-implementing.
    • Tautulli

      • Part of the Plex ecosystem which I abandoned. It was useful software, but unfortunately locked to Plex.
    • Unmanic

      • Never really had a use for this, though I thought I would at the time.
    • TT-RSS

      • Project is decent but the author is a asshole and very user-hostile, so I dropped it when I retooled my homelab a few years ago.
    • Jackett

      • Supplanted with Prowlarr.
    • Ombi

      • Same as the reason for dropping Overseerr.
    • Neko

      • Did it’s job, consumed lots of resources, and no arm64 docker image; though I managed to build my own. Got rid of it when I no longer had a use for the service.

    Archival/Documentation:

    • Filestash

      • Something about the project rubbed me the wrong way, vague on the details though.
    • Shiori

      • Was decent but the lack of updates then subsequent maintainer turnover scared me off. I check in from time to time to see how the project’s going.
    • Wallabag

      • Ended up being too slow and clunky for me, but that could be the hardware I was running it on at the time.
    • Archivebox

      • Same as above, but it definitely wasn’t the hardware.
    • Bookstack

      • It was alright but decided I didn’t need a separate service for documentation, I just use Code-Server with a documentation repo and raneto to give me a pretty page to navigate and for the family.
    • Filerun

      • Worked well while I was using it, not a fan of the closed-source nature and just didn’t feel the need to redeploy when I retooled my infrastructure.
    • Wiki.js

      • Same as bookstack, didn’t really have a use for a separate service.

    Dashboards:

    I’m going to preface this by saying I have some sort of addiction with dashboards, it’s unhealthy really.

    • Organizr

      • Didn’t like how everything was an iframe and it seemed particularly resource heavy for what I needed it to do.
    • Heimdall

      • My second dashboard. Liked the API integration not so much the design.
    • Homer

      • Wasn’t a fan of the design.
    • Homarr

      • Also didn’t like the design much.
    • Flame

      • Decent project, but decided to move on to something configuration file based.
    • Sui

      • Liked this one a lot and used for quite a while before homepage lured me away with API widgets.

    Infrastructure:

    • Apt-cacher-ng

      • Inadvertently made my infrastructure brittle with how I had it implemented. Decided to just rebuild my cluster’s cloud image on-demand instead of daily and update my apt distros the old-fashioned way.
    • LLDAP

      • Liked this project a lot, but added complexity to my infrastructure that could be more simply achieved other ways.
    • OpenLDAP

      • Same as lldap, but more feature rich and thus even more complicated.
    • Docker Registry

      • Set-up briefly but found a better use-case with Gitea’s integrated package registry which I’d already had deployed.
    • Guacamole

      • Used this for a while, but the clipboard situation sucked at the time and I gravitated to just using SSH anyway, and since I have Proxmox on hypervisor duties just used xterm.js or noVNC for console access.
    • Watchtower

      • Did it’s job but the :latest tag is dangerous to use. I like having change logs, an evaluation environment, and an approval based update workflow so I switched to renovate-bot.
    • Netmaker

      • Was a decent option for sure and faster than what I’m using currently in theory, but seemed a little to unnecessarily complicated to keep running for me.
    • Netboot.xyz

      • Definitely useful. Will probably redeploy it at some point.
    • OpenSSH-Server

      • Supplanted with Wireguard implementation in Tailscale.
    • Node-Red

      • Tried it very briefly but N8N fit my use-case better.