… But consolidation subtly changes the meaning of reliability. Before, if one Raspberry Pi died or one small server went sideways, only one service usually went down with it. Once several services share a single Proxmox host, the blast radius gets much bigger. …
… For enthusiasts running larger smart homes, local DNS services can provide another layer of reliability. …
… That flexibility also changes how you think about reliability. When the same machine hosts multiple small services, you’re not chasing performance for bragging rights. …
… Everything that was Docker deployable and self-hostable on my own server was much better than a premium cloud service. Ntfy had native support on my current monitoring service, Uptime Kuma. …
… Every restart broke my automations, every update interrupted services, and nothing ever felt stable. …
… With this foundation, I started replacing paid services that I used regularly with self-hosted alternatives. Immich replaced Google Photos for me AI-powered backups without the subscription Google Photos and Apple Photos were the first services that I wanted to replace. …
… It needs to solve a clear pain point, reduce maintenance, improve reliability, or unlock something I genuinely plan to use. If it doesn’t do one of those things, it probably belongs on a test box, not in the environment that handles daily services. …
… There’s no dependency on cloud services and no usage anxiety. …
… I have hosted many services, such as Immich, Jellyfin, and Nextcloud, on my home server, and Uptime Kuma for the uptime monitoring of these services on the same server. Everything looked alright, a few self-hosted services and a monitoring tool to watch over them. …
… None of those choices is wrong, but together they create a stack you have to understand when something breaks.The issue is not the number of services by itself. It’s the number of services that depend on other services before you can even reach the thing you need. …