A practical BorgBackup setup for NixOS with rsync.net: the manual CLI basics first, then a clean declarative `services.borgbackup.jobs` configuration with `sops-nix`, retention, pre-backup dump chaining, and restore verification.
A host-centric dotfiles repo is easy to start and easy to outgrow. Here's what improved when I reorganized mine into a Dendritic-style layout around reusable aspects, public vs private modules, and shared inventory — plus the tradeoffs in indirection, naming, and boilerplate.
One flake can handle both NixOS and plain Debian or Ubuntu deploys. Here's a practical pattern for embedding a full legacy deployment workflow into flake apps: setup, install, build, systemd user units, database dumps, REPL shells, and shell aliases.
Why the existing Python and Go gpsd exporters keep crashing, how a Rust rewrite fixes it, and wiring the whole thing up with Prometheus on NixOS — including the gpsd configuration that nobody tells you about.
Building hickory-dns from source with Nix, generating zone files from a host registry, wiring up DDNS with Kea DHCP, hardening with systemd, and monitoring the whole thing with Prometheus and Grafana.
How to install NixOS on cheap VPS instances with 768MB–1GB RAM using nixos-infect and a custom kexec image for nixos-anywhere.
Building your own private Cachix-style Nix binary cache on NixOS with Attic, Garage, PostgreSQL, sops-nix, multi-tenant token scoping, and GitHub Actions.
How to use a Nix flake to pin the Android SDK, wrap Flutter release builds, keep signing material out of the repo, and add idempotent emulator helpers.
Getting Ecowitt weather station data into Prometheus and Grafana with a Rust exporter and a NixOS module — with forwarding to Home Assistant and other receivers.
Using Crane to get reproducible Nix builds for Rust while keeping the standard Cargo workflow intact for developers who don't use Nix.