feat(vaultwarden): add SMTP configuration options and enhance signup settings

- Introduced SMTP settings for Vaultwarden including host, port, security, and authentication details.
- Added variables for signup verification, 2FA settings, password hints, and logging options.
- Updated Vaultwarden deployment to utilize new SMTP configurations.
- Enhanced Grafana module to support dynamic dashboard and datasource provisioning.
- Added LLM proxy configuration for Open Web UI with necessary environment variables.

ansible/host_vars/gpu-01.lab.alexpires.me.yml

@@ -88,8 +88,6 @@ duo_users: ["*", "!provision", "!comfyui", "!kokoro", "!ollama"]
 duo_api_hostname: api-7cda1654.duosecurity.com
 
 # podman services settings
-ollama_home: "/mnt/data/ollama"
-llama_home: "/mnt/data/llamacpp"
 podman_user_pubkey: "ecdsa-sha2-nistp521 AAAAE2VjZHNhLXNoYTItbmlzdHA1MjEAAAAIbmlzdHA1MjEAAACFBAHHOPBR9p9kq5Cqzpe4cr3jHnweaYrHPXv5sXNzt+sCXP54uc5rWUBhxW2OQVvQzJ47dEVhEKi4WSC7LcuKS2G5AQDzWXNiasHvYIYQU3F/EknVCZnsiXYqXphYkJA6rJCNRnISZCIC1mocq6PB7J08ONdRFCvjfUBuVRT8BNGXNmQ/zQ=="
 
 # ComfyUI settings
@@ -108,39 +106,18 @@ certbot_domains:
   - "gpu-01.lab.alexpires.me"
 
 certbot_cockpit_domain: "gpu-01.lab.alexpires.me"
-ollama_models:
-  # Thinking models
-  - gpt-oss:latest
-  - gemma4:e4b
-  - qwen3.5:9b
-  - qwen3.5:4b
-  # Non-thinking models
-  - ministral-3:3b
-  - ministral-3:8b
-  - ministral-3:14b
-  - granite4.1:3b
-  - granite4.1:8b
-  - deepseek-r1:latest
-  - qwen2.5-coder:3b
-  # Embedding models
-  - embeddinggemma:latest
-  - nomic-embed-text:latest
-  # Whisper models
-  - dimavz/whisper-tiny:latest
-ollama_model_files:
-  - file: "ollama/gpt-oss-latest-32k.modelfile"
-    name: "gpt-oss-32k:latest"
-  - file: "ollama/gemma4-e4b-32k.modelfile"
-    name: "gemma4-e4b-32k:latest"
-  - file: "ollama/qwen3.5-9b-32k.modelfile"
-    name: "qwen3.5-9b-32k:latest"
-  - file: "ollama/qwen3.5-9b-64k.modelfile"
-    name: "qwen3.5-9b-64k:latest"
-  - file: "ollama/qwen3.5-4b-32k.modelfile"
-    name: "qwen3.5-4b-32k:latest"
-  - file: "ollama/qwen3.5-4b-64k.modelfile"
-    name: "qwen3.5-4b-64k:latest"
-llama_models:
+
+# Llama.cpp server settings
+llama_server_home: "/mnt/data/llamacpp"
+llama_server_user_pubkey: "ecdsa-sha2-nistp521 AAAAE2VjZHNhLXNoYTItbmlzdHA1MjEAAAAIbmlzdHA1MjEAAACFBAHHOPBR9p9kq5Cqzpe4cr3jHnweaYrHPXv5sXNzt+sCXP54uc5rWUBhxW2OQVvQzJ47dEVhEKi4WSC7LcuKS2G5AQDzWXNiasHvYIYQU3F/EknVCZnsiXYqXphYkJA6rJCNRnISZCIC1mocq6PB7J08ONdRFCvjfUBuVRT8BNGXNmQ/zQ=="
+
+llama_server_port: 11434
+llama_server_devices:
+  - "nvidia.com/gpu=all"
+  - "/dev/kfd:/dev/kfd:rw"
+  - "/dev/dri:/dev/dri:rw"
+
+llama_server_models:
   # TODO: replace revision values with pinned commit SHAs for strict reproducibility.
   - name: "gpt-oss-20b-q4-0"
     model_id: "unsloth/gpt-oss-20b-GGUF"
@@ -186,9 +163,13 @@ llama_models:
     model_id: "unsloth/Devstral-Small-2-24B-Instruct-2512-GGUF"
     quant: "IQ4_XS"
     revision: "main"
+  - name: "qwen3-6-27b-mtp-iq4-xs"
+    model_id: "unsloth/Qwen3.6-27B-MTP-GGUF"
+    quant: "IQ4_XS"
+    revision: "main"
 
-# Optional global model preset options applied under [*] in models.ini.
-# Keys map to llama.cpp CLI args without leading dashes, short args, or LLAMA_ARG_* env names.
-llama_preset_global:
-  c: 131072
-  n-gpu-layers: all
+# # Optional global model preset options applied under [*] in models.ini.
+# # Keys map to llama.cpp CLI args without leading dashes, short args, or LLAMA_ARG_* env names.
+# llama_preset_global:
+#   c: 131072
+#   n-gpu-layers: all