alexpires.me/a13labs.infra
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Add GPU monitoring services and configurations

Browse Source 
- Create systemd service templates for llama_exporter, nvidia_exporter, and podman.
- Add Prometheus configuration template for GPU metrics scraping.
- Introduce variables for GPU monitoring role in main.yml.
- Implement tasks for syncing llama models, including user setup and package installation.
- Update podman tasks to ensure proper sudo access and lingering for the podman user.
- Modify inventory to include gpu_monitoring group.
- Add Terraform modules for deploying Vaultwarden, including ingress and service configurations.
- Create Grafana dashboard for real-time GPU monitoring metrics.
- Update secrets and environment files to include Vaultwarden admin token.
This commit is contained in:
Alexandre Pires
2026-05-29 23:42:14 +02:00
parent e9e28a5ca8
commit 350650ecc2
47 changed files with 3727 additions and 49 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ansible/host_vars/gpu-01.lab.alexpires.me.yml
+54 -1
View File
@@ -47,7 +47,8 @@ fw_allowed_ports:
- { rule: "allow", port: "8999", proto: "tcp", from: "10.19.4.0/24" } # ComfyUI - SimpleHttpServer (local network)
- { rule: "allow", port: "8012", proto: "tcp", from: "10.19.4.0/24" } # Qwen2 (local network)
- { rule: "allow", port: "8012", proto: "tcp", from: "10.5.5.5/32" } # Qwen2 (Laptop)
- { rule: "allow", port: "9090", proto: "tcp", from: "10.19.4.0/24" } # local network
- { rule: "allow", port: "9091", proto: "tcp", from: "10.19.4.0/24" } # Prometheus (local network)
- { rule: "allow", port: "9091", proto: "tcp", from: "10.5.5.5/32" } # Prometheus (Laptop)
ufw_outgoing_traffic:
- 22 # SSH
- 53 # DNS
@@ -139,3 +140,55 @@ ollama_model_files:
name: "qwen3.5-4b-32k:latest"
- file: "ollama/qwen3.5-4b-64k.modelfile"
name: "qwen3.5-4b-64k:latest"
llama_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"
quant: "Q4_0"
revision: "main"
- name: "gemma-4-26b-a4b-ud-iq4-xs"
model_id: "unsloth/gemma-4-26B-A4B-it-GGUF"
quant: "UD-IQ4_XS"
revision: "main"
- name: "gemma-4-31b-iq4-xs"
model_id: "unsloth/gemma-4-31B-it-GGUF"
quant: "IQ4_XS"
revision: "main"
- name: "qwen3-6-35b-a3b-mtp-ud-iq4-xs"
model_id: "unsloth/Qwen3.6-35B-A3B-MTP-GGUF"
quant: "UD-IQ4_XS"
revision: "main"
- name: "qwen3-5-27b-mtp-iq4-xs"
model_id: "unsloth/Qwen3.5-27B-MTP-GGUF"
quant: "IQ4_XS"
revision: "main"
- name: "qwen3-5-35b-a3b-ud-iq4-xs"
model_id: "unsloth/Qwen3.5-35B-A3B-GGUF"
quant: "UD-IQ4_XS"
revision: "main"
- name: "ministral-3-14b-instruct-iq4-xs"
model_id: "unsloth/Ministral-3-14B-Instruct-2512-GGUF"
quant: "IQ4_XS"
revision: "main"
- name: "granite-4-1-30b-iq4-xs"
model_id: "unsloth/granite-4.1-30b-GGUF"
quant: "IQ4_XS"
revision: "main"
- name: "qwen3-5-9b-iq4-xs"
model_id: "unsloth/Qwen3.5-9B-GGUF"
quant: "IQ4_XS"
revision: "main"
- name: "qwen3-6-35b-a3b-claude47-distilled-iq4-xs"
model_id: "lordx64/Qwen3.6-35B-A3B-Claude-4.7-Opus-Reasoning-Distilled-IQ4_XS-GGUF"
quant: "IQ4_XS"
revision: "main"
- name: "devstral-small-2-24b-instruct-iq4-xs"
model_id: "unsloth/Devstral-Small-2-24B-Instruct-2512-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