mirror of
https://github.com/Alkatrazz24/Funk-lab.git
synced 2026-07-09 01:34:41 +02:00
ALI YESILKAYA
470c69881c
Trois instances llama-server CPU manuelles (hors Ansible) tournaient sur gpu-01
sans servir : déjà orphelines de LiteLLM (alias absents de config.yaml, d'où le
bruit "Invalid model name"). Suppression complète :
gpu-01 (opérationnel, déjà fait) :
- systemctl disable --now + rm des units llama-server-{system,monitor,dev}
- ports firewall 1235/1236/1237 fermés
Repo :
- monitoring : scrape jobs (values.yaml) + alertes LlamaServer{System,Monitor}Down
(alerts-ai.yaml) retirés → plus de "target down"/alertes fantômes ; panneaux de
statut dashboard-infrastructure recyclés (system→llama-embed, monitor supprimé).
- docs admin/ + hermes-skills (souls funk-ai/brain, SKILL agent-delegation) mises à
jour : inventaire modèles, profils Hermès system/monitor sans backend (à repointer
sur qwen3-8b), litellm, ops, monitoring.
Les profils Hermès system/monitor existent toujours mais sans backend dédié.
Dashboards Grafana dashboard-ai (panneaux graphiques CPU) laissés tels quels
(cosmétique, séries vides, aucune alerte).
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
161 lines
3.2 KiB
YAML
161 lines
3.2 KiB
YAML
# kube-prometheus-stack — cluster Funk
|
|
# Ressources limitées pour compute-02/03 (8 GB RAM)
|
|
|
|
grafana:
|
|
initChownData:
|
|
enabled: false
|
|
grafana.ini:
|
|
database:
|
|
type: postgres
|
|
host: 192.168.10.1:5432
|
|
name: grafana
|
|
user: grafana
|
|
envFromSecrets:
|
|
- name: grafana-db-secret
|
|
ingress:
|
|
enabled: true
|
|
ingressClassName: traefik
|
|
hosts:
|
|
- grafana.lab.local
|
|
adminPassword: "funk-grafana"
|
|
resources:
|
|
requests:
|
|
cpu: 100m
|
|
memory: 256Mi
|
|
limits:
|
|
memory: 1Gi
|
|
persistence:
|
|
enabled: true
|
|
storageClassName: nfs
|
|
size: 2Gi
|
|
|
|
prometheus:
|
|
ingress:
|
|
enabled: true
|
|
ingressClassName: traefik
|
|
hosts:
|
|
- prometheus.lab.local
|
|
paths:
|
|
- /
|
|
pathType: Prefix
|
|
|
|
prometheusSpec:
|
|
retention: 15d
|
|
retentionSize: "8GB"
|
|
resources:
|
|
requests:
|
|
cpu: 200m
|
|
memory: 512Mi
|
|
limits:
|
|
memory: 1Gi
|
|
storageSpec:
|
|
volumeClaimTemplate:
|
|
spec:
|
|
storageClassName: nfs
|
|
accessModes: ["ReadWriteOnce"]
|
|
resources:
|
|
requests:
|
|
storage: 20Gi
|
|
|
|
# Scrape des nœuds hors cluster (storage-01, gpu-01)
|
|
additionalScrapeConfigs:
|
|
- job_name: storage-01
|
|
static_configs:
|
|
- targets: ['192.168.10.1:9100']
|
|
labels:
|
|
instance: storage-01
|
|
role: gateway
|
|
|
|
- job_name: gpu-01-node
|
|
static_configs:
|
|
- targets: ['192.168.10.20:9100']
|
|
labels:
|
|
instance: gpu-01
|
|
role: gpu
|
|
|
|
- job_name: llama-server-gpu
|
|
static_configs:
|
|
- targets: ['192.168.10.20:1234']
|
|
labels:
|
|
instance: gpu-01
|
|
model: qwen3-8b
|
|
type: gpu
|
|
|
|
alertmanager:
|
|
ingress:
|
|
enabled: true
|
|
ingressClassName: traefik
|
|
hosts:
|
|
- alertmanager.lab.local
|
|
paths:
|
|
- /
|
|
pathType: Prefix
|
|
|
|
alertmanagerSpec:
|
|
resources:
|
|
requests:
|
|
cpu: 50m
|
|
memory: 64Mi
|
|
limits:
|
|
memory: 128Mi
|
|
config:
|
|
global:
|
|
resolve_timeout: 5m
|
|
route:
|
|
group_by: ['alertname', 'instance']
|
|
group_wait: 30s
|
|
group_interval: 5m
|
|
repeat_interval: 4h
|
|
receiver: hermes-monitor
|
|
routes:
|
|
# Watchdog (heartbeat interne kube-prometheus-stack) — silencé
|
|
- matchers:
|
|
- alertname = "Watchdog"
|
|
receiver: "null"
|
|
repeat_interval: 24h
|
|
receivers:
|
|
- name: "null"
|
|
- name: hermes-monitor
|
|
webhook_configs:
|
|
- url: 'http://192.168.10.1:9093/webhook'
|
|
send_resolved: true
|
|
- url: 'http://n8n.lab.local/webhook/alerts'
|
|
send_resolved: true
|
|
|
|
prometheusOperator:
|
|
resources:
|
|
requests:
|
|
cpu: 100m
|
|
memory: 128Mi
|
|
limits:
|
|
memory: 256Mi
|
|
|
|
kube-state-metrics:
|
|
resources:
|
|
requests:
|
|
cpu: 50m
|
|
memory: 64Mi
|
|
limits:
|
|
memory: 128Mi
|
|
|
|
nodeExporter:
|
|
resources:
|
|
requests:
|
|
cpu: 50m
|
|
memory: 32Mi
|
|
limits:
|
|
memory: 64Mi
|
|
|
|
prometheus-node-exporter:
|
|
extraArgs:
|
|
- --collector.rapl
|
|
|
|
# Désactiver etcd/scheduler monitoring (Talos — non accessible)
|
|
kubeEtcd:
|
|
enabled: false
|
|
kubeScheduler:
|
|
enabled: false
|
|
kubeControllerManager:
|
|
enabled: false
|
|
kubeProxy:
|
|
enabled: false
|