I plan to have the following services running concurrently on it:

  • A VPN (OpenVPN or Wireguard)
  • A very lightweight personal website
  • A Nextcloud instance (25GB storage max)
  • A Vaultwarden instance
  • An Invidious instance
  • A Matrix server
  • A Lemmy instance

I’m unsure if these would be private or public instances. But I’d be curious to hear any thoughts on how much more space I’d need for public instances too, if you’d have a sense of that.

I currently have a VPS with 2GB RAM + 50GB storage. Would that be enough? Thanks in advance!

    • mrmojo@beehaw.org
      link
      fedilink
      arrow-up
      2
      ·
      1 year ago

      You could also check out Conduit, it’s still in beta but very lightweight.

      Conduit is a lightweight open-source server implementation of the Matrix Specification with a focus on easy setup and low system requirements. That means you can make your own Conduit setup in just a few minutes.

    • cnschn@lemmy.cnschn.com
      link
      fedilink
      arrow-up
      1
      ·
      1 year ago

      So true, every few months I get the urge to ditch Synapse, but then I think about how much of a hassle it would be to migrate and happily pay a few bucks more for the extra RAM on the VPS. Don’t be me, use Dendrite right from the start.

  • Justin@lemmy.jlh.name
    link
    fedilink
    arrow-up
    0
    ·
    1 year ago

    The heaviest things will probably be the postgres database(s) needed for nextcloud, vaultwarden, and lemmy. (not sure if invidious or matrix use dbs). I would say 500-1000MiB per DB, so probably around 4GiB to be safe.

    • Trainguyrom@reddthat.com
      link
      fedilink
      arrow-up
      1
      ·
      1 year ago

      Can’t you have multiple services using the same database? My understanding is that they would each do different tables and if you give each a different database user that’s not only good practice but also helps prevent them from stomping on eachother

      • cnschn@lemmy.cnschn.com
        link
        fedilink
        arrow-up
        3
        ·
        1 year ago

        You could, but it probably wouldn’t help much. The overhead for each additional Postgres server is minimal, the RAM usage comes from each database. It doesn’t really matter if those are on the same postgres instance or on separate ones, in my experience.

  • Rick@thesimplecorner.org
    link
    fedilink
    arrow-up
    0
    ·
    1 year ago

    I am running Lemmy on mine with 1gb ram. 1 cpu. Ubuntu 22.04 server. I almost run out of ram. Seen as low as 50mb. Fwiw. My experience.

    Probably will bump mine up.

    Also, I found out my VPS host doesn’t allow smtp traffic (25 or 465)… Anyone got suggestions on hosts that are cheap, reputable that allow smtp?

  • poVoq@slrpnk.net
    link
    fedilink
    arrow-up
    0
    ·
    1 year ago

    2gb will be probably insufficient, mainly because of Matrix Synapse. If you avoid joining any large communities it might work, but if you are not doing that you might as well use a XMPP server which is better anyways.

    A small Lemmy instance will take around 500-1000mb RAM so that will likely work but use a significant portion of your RAM.

      • darkfoe@lemmy.serverfail.party
        link
        fedilink
        arrow-up
        2
        ·
        edit-2
        1 year ago

        This is my droplet with 1GB of RAM only running lemmy:

        free -m
                       total        used        free      shared  buff/cache   available
        Mem:             964         386          68         141         509         219
        Swap:           2047         310        1737
        

        So expect at least 1GB for lemmy with postgres included when you include spikes etc.