- Updated `cudaPackages` to `pkgs.cudaPackages_12_3`
- Changed `boost` version from `pkgs.unstablePkgs.boost186` to `pkgs.boost186`
- Disabled `pulseaudio.enable` and removed redundant `sound.enable`
- Added `redlib` systemd service environment variables and settings
- Enabled `redlib` service with specified port and package
- Configured virtual host for `reddit.giugl.io` to forward requests to the `redlib` service
- Updated `dns_config` to `dns` and adjusted nested fields accordingly.
- Commented out default log level and base domain for clarity.
- Changed `ip_prefixes` to `prefixes.v4` to match expected configuration format.
- Removed unused frontend domain, port, and listen address variables
- Centralized `ollamaHost` and `ollamaPort` usage
- Cleaned up commented-out sections for future reference or removal
- Ensured consistent placement of `acceleration` in the `services.ollama` block
- Added `teslamate-flake` input to the flake configuration
- Imported TeslaMate's NixOS module in the system configuration
- Updated function signatures to remove unused `system` parameter where applicable
- Introduced `recommendedProxySettings` to force the use of recommended proxy configuration.
- Updated location configuration to include `recommendedProxySettings`.
- Updated `domain` to use a single string "vipienne.giugl.io"
- Changed `headscalePkg` to use the stable `pkgs.headscale` instead of `pkgs.unstablePkgs.headscale`
- Corrected `base_domain` in `dns_config` to use the updated `domain` variable
- Commented out the `frontendDomain` vhost configuration
- Updated backend port to use `ollamaPort`
- Added `recommendedProxySettings = false;` for backend vhost
- Commented out the entire OCI containers section for `ollama-webui`
- Updated domain to `tesla.giugl.io`
- Added separate ports for TeslaMate (`11234`) and Grafana (`11334`)
- Configured proxy settings for websockets in `/` and `/live/websocket` locations
- Added `/grafana` location with appropriate configuration
- Updated `services.teslamate` to include port, listen address, secrets file, virtual host, PostgreSQL server, and Grafana settings
- Removed inline `config` and `roles` definitions
- Used `lib.readFile` to load external files for configuration and roles
- Organized role files into separate markdown documents
- Updated `type` to `openai-compatible`
- Added `name` field for clarity
- Adjusted `api_base` URL to include `/v1`
- Simplified and updated model configurations for consistency
