diff --git a/argocd/aerodrome/prometheus-app.yaml b/argocd/aerodrome/prometheus-app.yaml
new file mode 100644
index 0000000..40dd927
--- /dev/null
+++ b/argocd/aerodrome/prometheus-app.yaml
@@ -0,0 +1,27 @@
+apiVersion: argoproj.io/v1alpha1
+kind: Application
+metadata:
+  name: aerodrome-prometheus
+  namespace: infrastructure
+  labels:
+    app.kubernetes.io/name: aerodrome-prometheus
+    cluster: aerodrome
+  annotations:
+    notifications.argoproj.io/subscribe.on-sync-failed.gotify: ""
+    notifications.argoproj.io/subscribe.on-health-degraded.gotify: ""
+spec:
+  project: homelab
+  source:
+    repoURL: "https://git.snarfnet.net/Snarf/homelab-gitops.git"
+    targetRevision: HEAD
+    path: manifests/aerodrome/prometheus
+  destination:
+    server: "https://140.44.4.11:6443"
+    namespace: default
+  syncPolicy:
+    automated:
+      prune: false
+      selfHeal: true
+    syncOptions:
+      - CreateNamespace=true
+      - ServerSideApply=true
diff --git a/manifests/aerodrome/grafana/ai-stack-dashboard.json b/manifests/aerodrome/grafana/ai-stack-dashboard.json
new file mode 100644
index 0000000..0d0e0ba
--- /dev/null
+++ b/manifests/aerodrome/grafana/ai-stack-dashboard.json
@@ -0,0 +1,523 @@
+{
+  "annotations": {
+    "list": [
+      {
+        "builtIn": 1,
+        "datasource": "-- Grafana --",
+        "enable": true,
+        "hide": true,
+        "name": "Annotations & Alerts",
+        "type": "dashboard"
+      }
+    ]
+  },
+  "editable": true,
+  "fiscalYearStartMonth": 0,
+  "graphTooltip": 1,
+  "id": null,
+  "links": [],
+  "liveNow": false,
+  "panels": [
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 0
+      },
+      "id": 100,
+      "title": "\u26a1 AI Stack Executive Overview & Cloud Cost Offsets",
+      "type": "row"
+    },
+    {
+      "id": 1,
+      "title": "Prompt Tokens Processed (In)",
+      "type": "stat",
+      "gridPos": {
+        "h": 4,
+        "w": 6,
+        "x": 0,
+        "y": 1
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "sum(vllm:prompt_tokens_total)",
+          "legendFormat": "Prompt Tokens",
+          "refId": "A"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "blue",
+                "value": null
+              }
+            ]
+          },
+          "unit": "locale"
+        }
+      },
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "center"
+      }
+    },
+    {
+      "id": 2,
+      "title": "Generation Tokens Synthesized (Out)",
+      "type": "stat",
+      "gridPos": {
+        "h": 4,
+        "w": 6,
+        "x": 6,
+        "y": 1
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "sum(vllm:generation_tokens_total)",
+          "legendFormat": "Gen Tokens",
+          "refId": "A"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "locale"
+        }
+      },
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "center"
+      }
+    },
+    {
+      "id": 3,
+      "title": "Homelab Cloud API Savings ($USD)",
+      "type": "stat",
+      "gridPos": {
+        "h": 4,
+        "w": 6,
+        "x": 12,
+        "y": 1
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "(sum(vllm:prompt_tokens_total) * 0.00000025) + (sum(vllm:generation_tokens_total) * 0.00000100)",
+          "legendFormat": "Net Savings",
+          "refId": "A"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "currencyUSD"
+        }
+      },
+      "options": {
+        "colorMode": "background",
+        "graphMode": "none",
+        "justifyMode": "center"
+      }
+    },
+    {
+      "id": 4,
+      "title": "Active GPU Cluster Health",
+      "type": "stat",
+      "gridPos": {
+        "h": 4,
+        "w": 6,
+        "x": 18,
+        "y": 1
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "count(DCGM_FI_DEV_GPU_TEMP)",
+          "legendFormat": "Tesla T4s Active",
+          "refId": "A"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "purple",
+                "value": null
+              }
+            ]
+          },
+          "unit": "short"
+        }
+      },
+      "options": {
+        "colorMode": "value",
+        "graphMode": "none",
+        "justifyMode": "center"
+      }
+    },
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 5
+      },
+      "id": 200,
+      "title": "\ud83c\udfae Tesla T4 GPU Cluster Telemetry (Harvester HCI Spacedock)",
+      "type": "row"
+    },
+    {
+      "id": 5,
+      "title": "GPU Core Utilization (%)",
+      "type": "timeseries",
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 6
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "DCGM_FI_DEV_GPU_UTIL",
+          "legendFormat": "{{Hostname}} ({{modelName}})",
+          "refId": "A"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "drawStyle": "line",
+            "lineInterpolation": "smooth",
+            "fillOpacity": 20
+          },
+          "unit": "percent",
+          "min": 0,
+          "max": 100
+        }
+      }
+    },
+    {
+      "id": 6,
+      "title": "VRAM Memory Allocated & Used",
+      "type": "timeseries",
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 6
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "DCGM_FI_DEV_FB_USED * 1024 * 1024",
+          "legendFormat": "{{Hostname}} VRAM Used",
+          "refId": "A"
+        },
+        {
+          "expr": "(DCGM_FI_DEV_FB_USED + DCGM_FI_DEV_FB_FREE) * 1024 * 1024",
+          "legendFormat": "{{Hostname}} VRAM Total",
+          "refId": "B"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "drawStyle": "line",
+            "lineInterpolation": "smooth",
+            "fillOpacity": 15
+          },
+          "unit": "bytes"
+        }
+      }
+    },
+    {
+      "id": 7,
+      "title": "GPU Thermal Profile (\u00b0C)",
+      "type": "timeseries",
+      "gridPos": {
+        "h": 7,
+        "w": 12,
+        "x": 0,
+        "y": 14
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "DCGM_FI_DEV_GPU_TEMP",
+          "legendFormat": "{{Hostname}} (\u00b0C)",
+          "refId": "A"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "drawStyle": "line",
+            "lineInterpolation": "smooth",
+            "fillOpacity": 10
+          },
+          "unit": "celsius",
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "yellow",
+                "value": 65
+              },
+              {
+                "color": "red",
+                "value": 80
+              }
+            ]
+          }
+        }
+      }
+    },
+    {
+      "id": 8,
+      "title": "GPU Power Draw (Watts)",
+      "type": "timeseries",
+      "gridPos": {
+        "h": 7,
+        "w": 12,
+        "x": 12,
+        "y": 14
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "DCGM_FI_DEV_POWER_USAGE",
+          "legendFormat": "{{Hostname}} (Watts)",
+          "refId": "A"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "drawStyle": "line",
+            "lineInterpolation": "smooth",
+            "fillOpacity": 15
+          },
+          "unit": "watt"
+        }
+      }
+    },
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 21
+      },
+      "id": 300,
+      "title": "\ud83e\udde0 vLLM Real-Time Inference Dynamics & Token Velocity",
+      "type": "row"
+    },
+    {
+      "id": 9,
+      "title": "Token Generation Velocity (Tokens / Sec)",
+      "type": "timeseries",
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 22
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "sum(rate(vllm:generation_tokens_total[2m])) by (model_name)",
+          "legendFormat": "{{model_name}} (Out)",
+          "refId": "A"
+        },
+        {
+          "expr": "sum(rate(vllm:prompt_tokens_total[2m])) by (model_name)",
+          "legendFormat": "{{model_name}} (In)",
+          "refId": "B"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "drawStyle": "line",
+            "lineInterpolation": "smooth",
+            "fillOpacity": 25
+          },
+          "unit": "cps"
+        }
+      }
+    },
+    {
+      "id": 10,
+      "title": "Inference Concurrency & Request Backlog",
+      "type": "timeseries",
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 22
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "vllm:num_requests_running",
+          "legendFormat": "{{model_name}} Running",
+          "refId": "A"
+        },
+        {
+          "expr": "vllm:num_requests_waiting",
+          "legendFormat": "{{model_name}} Queued",
+          "refId": "B"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "drawStyle": "line",
+            "lineInterpolation": "stepAfter",
+            "fillOpacity": 15
+          },
+          "unit": "short"
+        }
+      }
+    },
+    {
+      "id": 11,
+      "title": "GPU KV Cache Memory Saturation",
+      "type": "timeseries",
+      "gridPos": {
+        "h": 7,
+        "w": 12,
+        "x": 0,
+        "y": 30
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "vllm:gpu_cache_usage_factor * 100",
+          "legendFormat": "{{model_name}} Cache Usage",
+          "refId": "A"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "drawStyle": "line",
+            "lineInterpolation": "smooth",
+            "fillOpacity": 30
+          },
+          "unit": "percent",
+          "min": 0,
+          "max": 100
+        }
+      }
+    },
+    {
+      "id": 12,
+      "title": "NVIDIA Triton & Qdrant Operations",
+      "type": "timeseries",
+      "gridPos": {
+        "h": 7,
+        "w": 12,
+        "x": 12,
+        "y": 30
+      },
+      "datasource": "Prometheus",
+      "targets": [
+        {
+          "expr": "rate(nv_inference_request_success[2m])",
+          "legendFormat": "Triton Success/s",
+          "refId": "A"
+        },
+        {
+          "expr": "cpu_cores_used",
+          "legendFormat": "Qdrant CPU Cores",
+          "refId": "B"
+        }
+      ],
+      "fieldConfig": {
+        "defaults": {
+          "custom": {
+            "drawStyle": "line",
+            "lineInterpolation": "smooth"
+          },
+          "unit": "short"
+        }
+      }
+    }
+  ],
+  "refresh": "10s",
+  "schemaVersion": 39,
+  "tags": [
+    "ai",
+    "vllm",
+    "dcgm",
+    "gpu",
+    "snarfnet"
+  ],
+  "templating": {
+    "list": []
+  },
+  "time": {
+    "from": "now-6h",
+    "to": "now"
+  },
+  "timepicker": {
+    "refresh_intervals": [
+      "5s",
+      "10s",
+      "30s",
+      "1m",
+      "5m"
+    ]
+  },
+  "timezone": "browser",
+  "title": "SnarfNet AI Stack & Inference Intelligence",
+  "uid": "snarfnet-ai-stack",
+  "version": 1
+}
\ No newline at end of file
diff --git a/manifests/aerodrome/grafana/grafana.yaml b/manifests/aerodrome/grafana/grafana.yaml
index 990c54d..d2fab0e 100644
--- a/manifests/aerodrome/grafana/grafana.yaml
+++ b/manifests/aerodrome/grafana/grafana.yaml
@@ -111,6 +111,15 @@ spec:
           subPath: grafana.ini
         - mountPath: /var/lib/grafana
           name: storage
+        - mountPath: /etc/grafana/provisioning/datasources/datasources.yaml
+          name: datasources-config
+          subPath: datasources.yaml
+        - mountPath: /etc/grafana/provisioning/dashboards/dashboards.yaml
+          name: dashboards-provider-config
+          subPath: dashboards.yaml
+        - mountPath: /var/lib/grafana-dashboards/ai-stack-dashboard.json
+          name: ai-stack-dashboard
+          subPath: ai-stack-dashboard.json
       dnsPolicy: ClusterFirst
       enableServiceLinks: true
       initContainers:
@@ -154,6 +163,18 @@ spec:
       - name: storage
         persistentVolumeClaim:
           claimName: grafana
+      - configMap:
+          defaultMode: 420
+          name: grafana-datasources
+        name: datasources-config
+      - configMap:
+          defaultMode: 420
+          name: grafana-dashboards-provider
+        name: dashboards-provider-config
+      - configMap:
+          defaultMode: 420
+          name: grafana-dashboard-ai-stack
+        name: ai-stack-dashboard
 ---
 apiVersion: v1
 kind: Service
@@ -250,3 +271,575 @@ spec:
   - hosts:
     - grafana.snarfnet.net
     secretName: firefly-tls-cert
+---
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: grafana-datasources
+  namespace: default
+  labels:
+    app.kubernetes.io/name: grafana
+data:
+  datasources.yaml: |
+    apiVersion: 1
+    datasources:
+      - name: Prometheus
+        type: prometheus
+        access: proxy
+        url: http://prometheus:9090
+        isDefault: true
+        jsonData:
+          timeInterval: 15s
+---
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: grafana-dashboards-provider
+  namespace: default
+  labels:
+    app.kubernetes.io/name: grafana
+data:
+  dashboards.yaml: |
+    apiVersion: 1
+    providers:
+      - name: 'ai-stack-provider'
+        orgId: 1
+        folder: 'AI Stack & Inference'
+        type: file
+        disableDeletion: false
+        editable: true
+        options:
+          path: /var/lib/grafana-dashboards
+---
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: grafana-dashboard-ai-stack
+  namespace: default
+  labels:
+    app.kubernetes.io/name: grafana
+data:
+  ai-stack-dashboard.json: |
+    {
+      "annotations": {
+        "list": [
+          {
+            "builtIn": 1,
+            "datasource": "-- Grafana --",
+            "enable": true,
+            "hide": true,
+            "name": "Annotations & Alerts",
+            "type": "dashboard"
+          }
+        ]
+      },
+      "editable": true,
+      "fiscalYearStartMonth": 0,
+      "graphTooltip": 1,
+      "id": null,
+      "links": [],
+      "liveNow": false,
+      "panels": [
+        {
+          "collapsed": false,
+          "gridPos": {
+            "h": 1,
+            "w": 24,
+            "x": 0,
+            "y": 0
+          },
+          "id": 100,
+          "title": "\u26a1 AI Stack Executive Overview & Cloud Cost Offsets",
+          "type": "row"
+        },
+        {
+          "id": 1,
+          "title": "Prompt Tokens Processed (In)",
+          "type": "stat",
+          "gridPos": {
+            "h": 4,
+            "w": 6,
+            "x": 0,
+            "y": 1
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "sum(vllm:prompt_tokens_total)",
+              "legendFormat": "Prompt Tokens",
+              "refId": "A"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "color": {
+                "mode": "thresholds"
+              },
+              "thresholds": {
+                "mode": "absolute",
+                "steps": [
+                  {
+                    "color": "blue",
+                    "value": null
+                  }
+                ]
+              },
+              "unit": "locale"
+            }
+          },
+          "options": {
+            "colorMode": "value",
+            "graphMode": "area",
+            "justifyMode": "center"
+          }
+        },
+        {
+          "id": 2,
+          "title": "Generation Tokens Synthesized (Out)",
+          "type": "stat",
+          "gridPos": {
+            "h": 4,
+            "w": 6,
+            "x": 6,
+            "y": 1
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "sum(vllm:generation_tokens_total)",
+              "legendFormat": "Gen Tokens",
+              "refId": "A"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "color": {
+                "mode": "thresholds"
+              },
+              "thresholds": {
+                "mode": "absolute",
+                "steps": [
+                  {
+                    "color": "green",
+                    "value": null
+                  }
+                ]
+              },
+              "unit": "locale"
+            }
+          },
+          "options": {
+            "colorMode": "value",
+            "graphMode": "area",
+            "justifyMode": "center"
+          }
+        },
+        {
+          "id": 3,
+          "title": "Homelab Cloud API Savings ($USD)",
+          "type": "stat",
+          "gridPos": {
+            "h": 4,
+            "w": 6,
+            "x": 12,
+            "y": 1
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "(sum(vllm:prompt_tokens_total) * 0.00000025) + (sum(vllm:generation_tokens_total) * 0.00000100)",
+              "legendFormat": "Net Savings",
+              "refId": "A"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "color": {
+                "mode": "thresholds"
+              },
+              "thresholds": {
+                "mode": "absolute",
+                "steps": [
+                  {
+                    "color": "green",
+                    "value": null
+                  }
+                ]
+              },
+              "unit": "currencyUSD"
+            }
+          },
+          "options": {
+            "colorMode": "background",
+            "graphMode": "none",
+            "justifyMode": "center"
+          }
+        },
+        {
+          "id": 4,
+          "title": "Active GPU Cluster Health",
+          "type": "stat",
+          "gridPos": {
+            "h": 4,
+            "w": 6,
+            "x": 18,
+            "y": 1
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "count(DCGM_FI_DEV_GPU_TEMP)",
+              "legendFormat": "Tesla T4s Active",
+              "refId": "A"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "color": {
+                "mode": "thresholds"
+              },
+              "thresholds": {
+                "mode": "absolute",
+                "steps": [
+                  {
+                    "color": "purple",
+                    "value": null
+                  }
+                ]
+              },
+              "unit": "short"
+            }
+          },
+          "options": {
+            "colorMode": "value",
+            "graphMode": "none",
+            "justifyMode": "center"
+          }
+        },
+        {
+          "collapsed": false,
+          "gridPos": {
+            "h": 1,
+            "w": 24,
+            "x": 0,
+            "y": 5
+          },
+          "id": 200,
+          "title": "\ud83c\udfae Tesla T4 GPU Cluster Telemetry (Harvester HCI Spacedock)",
+          "type": "row"
+        },
+        {
+          "id": 5,
+          "title": "GPU Core Utilization (%)",
+          "type": "timeseries",
+          "gridPos": {
+            "h": 8,
+            "w": 12,
+            "x": 0,
+            "y": 6
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "DCGM_FI_DEV_GPU_UTIL",
+              "legendFormat": "{{Hostname}} ({{modelName}})",
+              "refId": "A"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "custom": {
+                "drawStyle": "line",
+                "lineInterpolation": "smooth",
+                "fillOpacity": 20
+              },
+              "unit": "percent",
+              "min": 0,
+              "max": 100
+            }
+          }
+        },
+        {
+          "id": 6,
+          "title": "VRAM Memory Allocated & Used",
+          "type": "timeseries",
+          "gridPos": {
+            "h": 8,
+            "w": 12,
+            "x": 12,
+            "y": 6
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "DCGM_FI_DEV_FB_USED * 1024 * 1024",
+              "legendFormat": "{{Hostname}} VRAM Used",
+              "refId": "A"
+            },
+            {
+              "expr": "(DCGM_FI_DEV_FB_USED + DCGM_FI_DEV_FB_FREE) * 1024 * 1024",
+              "legendFormat": "{{Hostname}} VRAM Total",
+              "refId": "B"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "custom": {
+                "drawStyle": "line",
+                "lineInterpolation": "smooth",
+                "fillOpacity": 15
+              },
+              "unit": "bytes"
+            }
+          }
+        },
+        {
+          "id": 7,
+          "title": "GPU Thermal Profile (\u00b0C)",
+          "type": "timeseries",
+          "gridPos": {
+            "h": 7,
+            "w": 12,
+            "x": 0,
+            "y": 14
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "DCGM_FI_DEV_GPU_TEMP",
+              "legendFormat": "{{Hostname}} (\u00b0C)",
+              "refId": "A"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "custom": {
+                "drawStyle": "line",
+                "lineInterpolation": "smooth",
+                "fillOpacity": 10
+              },
+              "unit": "celsius",
+              "thresholds": {
+                "mode": "absolute",
+                "steps": [
+                  {
+                    "color": "green",
+                    "value": null
+                  },
+                  {
+                    "color": "yellow",
+                    "value": 65
+                  },
+                  {
+                    "color": "red",
+                    "value": 80
+                  }
+                ]
+              }
+            }
+          }
+        },
+        {
+          "id": 8,
+          "title": "GPU Power Draw (Watts)",
+          "type": "timeseries",
+          "gridPos": {
+            "h": 7,
+            "w": 12,
+            "x": 12,
+            "y": 14
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "DCGM_FI_DEV_POWER_USAGE",
+              "legendFormat": "{{Hostname}} (Watts)",
+              "refId": "A"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "custom": {
+                "drawStyle": "line",
+                "lineInterpolation": "smooth",
+                "fillOpacity": 15
+              },
+              "unit": "watt"
+            }
+          }
+        },
+        {
+          "collapsed": false,
+          "gridPos": {
+            "h": 1,
+            "w": 24,
+            "x": 0,
+            "y": 21
+          },
+          "id": 300,
+          "title": "\ud83e\udde0 vLLM Real-Time Inference Dynamics & Token Velocity",
+          "type": "row"
+        },
+        {
+          "id": 9,
+          "title": "Token Generation Velocity (Tokens / Sec)",
+          "type": "timeseries",
+          "gridPos": {
+            "h": 8,
+            "w": 12,
+            "x": 0,
+            "y": 22
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "sum(rate(vllm:generation_tokens_total[2m])) by (model_name)",
+              "legendFormat": "{{model_name}} (Out)",
+              "refId": "A"
+            },
+            {
+              "expr": "sum(rate(vllm:prompt_tokens_total[2m])) by (model_name)",
+              "legendFormat": "{{model_name}} (In)",
+              "refId": "B"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "custom": {
+                "drawStyle": "line",
+                "lineInterpolation": "smooth",
+                "fillOpacity": 25
+              },
+              "unit": "cps"
+            }
+          }
+        },
+        {
+          "id": 10,
+          "title": "Inference Concurrency & Request Backlog",
+          "type": "timeseries",
+          "gridPos": {
+            "h": 8,
+            "w": 12,
+            "x": 12,
+            "y": 22
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "vllm:num_requests_running",
+              "legendFormat": "{{model_name}} Running",
+              "refId": "A"
+            },
+            {
+              "expr": "vllm:num_requests_waiting",
+              "legendFormat": "{{model_name}} Queued",
+              "refId": "B"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "custom": {
+                "drawStyle": "line",
+                "lineInterpolation": "stepAfter",
+                "fillOpacity": 15
+              },
+              "unit": "short"
+            }
+          }
+        },
+        {
+          "id": 11,
+          "title": "GPU KV Cache Memory Saturation",
+          "type": "timeseries",
+          "gridPos": {
+            "h": 7,
+            "w": 12,
+            "x": 0,
+            "y": 30
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "vllm:gpu_cache_usage_factor * 100",
+              "legendFormat": "{{model_name}} Cache Usage",
+              "refId": "A"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "custom": {
+                "drawStyle": "line",
+                "lineInterpolation": "smooth",
+                "fillOpacity": 30
+              },
+              "unit": "percent",
+              "min": 0,
+              "max": 100
+            }
+          }
+        },
+        {
+          "id": 12,
+          "title": "NVIDIA Triton & Qdrant Operations",
+          "type": "timeseries",
+          "gridPos": {
+            "h": 7,
+            "w": 12,
+            "x": 12,
+            "y": 30
+          },
+          "datasource": "Prometheus",
+          "targets": [
+            {
+              "expr": "rate(nv_inference_request_success[2m])",
+              "legendFormat": "Triton Success/s",
+              "refId": "A"
+            },
+            {
+              "expr": "cpu_cores_used",
+              "legendFormat": "Qdrant CPU Cores",
+              "refId": "B"
+            }
+          ],
+          "fieldConfig": {
+            "defaults": {
+              "custom": {
+                "drawStyle": "line",
+                "lineInterpolation": "smooth"
+              },
+              "unit": "short"
+            }
+          }
+        }
+      ],
+      "refresh": "10s",
+      "schemaVersion": 39,
+      "tags": [
+        "ai",
+        "vllm",
+        "dcgm",
+        "gpu",
+        "snarfnet"
+      ],
+      "templating": {
+        "list": []
+      },
+      "time": {
+        "from": "now-6h",
+        "to": "now"
+      },
+      "timepicker": {
+        "refresh_intervals": [
+          "5s",
+          "10s",
+          "30s",
+          "1m",
+          "5m"
+        ]
+      },
+      "timezone": "browser",
+      "title": "SnarfNet AI Stack & Inference Intelligence",
+      "uid": "snarfnet-ai-stack",
+      "version": 1
+    }
diff --git a/manifests/aerodrome/prometheus/prometheus.yaml b/manifests/aerodrome/prometheus/prometheus.yaml
new file mode 100644
index 0000000..fea60e5
--- /dev/null
+++ b/manifests/aerodrome/prometheus/prometheus.yaml
@@ -0,0 +1,182 @@
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: prometheus-storage
+  namespace: default
+  labels:
+    app.kubernetes.io/name: prometheus
+spec:
+  accessModes:
+    - ReadWriteOnce
+  storageClassName: longhorn
+  resources:
+    requests:
+      storage: 50Gi
+---
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: prometheus-config
+  namespace: default
+  labels:
+    app.kubernetes.io/name: prometheus
+data:
+  prometheus.yml: |
+    global:
+      scrape_interval: 15s
+      scrape_timeout: 10s
+      evaluation_interval: 15s
+
+    scrape_configs:
+      - job_name: 'prometheus'
+        static_configs:
+          - targets: ['localhost:9090']
+
+      - job_name: 'vllm-qwen'
+        metrics_path: /metrics
+        static_configs:
+          - targets: ['140.44.4.84:30800']
+            labels:
+              model: 'Qwen2.5-7B-Instruct'
+              node: 'snaisvr44hrvstr04'
+              role: 'general-instruction'
+
+      - job_name: 'vllm-gemma4'
+        metrics_path: /metrics
+        static_configs:
+          - targets: ['140.44.4.82:30804']
+            labels:
+              model: 'Gemma-4-12B-it'
+              node: 'snaisvr44hrvstr02'
+              role: 'coding-agent'
+
+      - job_name: 'vllm-deepseek-r1'
+        metrics_path: /metrics
+        static_configs:
+          - targets: ['140.44.4.81:30805']
+            labels:
+              model: 'DeepSeek-R1-Distill-Qwen-14B'
+              node: 'snaisvr44hrvstr01'
+              role: 'reasoning-planner'
+
+      - job_name: 'triton-inference-server'
+        metrics_path: /metrics
+        static_configs:
+          - targets: ['140.44.4.83:30803']
+            labels:
+              app: 'triton-inference-server'
+              node: 'snaisvr44hrvstr03'
+              model: 'decision_engine'
+
+      - job_name: 'qdrant'
+        metrics_path: /metrics
+        static_configs:
+          - targets: ['140.44.4.82:30633']
+            labels:
+              app: 'qdrant'
+              node: 'snaisvr44hrvstr02'
+
+      - job_name: 'nvidia-dcgm'
+        metrics_path: /metrics
+        static_configs:
+          - targets: ['140.44.4.81:9400']
+            labels:
+              node: 'snaisvr44hrvstr01'
+              gpu_model: 'Tesla T4'
+          - targets: ['140.44.4.82:9400']
+            labels:
+              node: 'snaisvr44hrvstr02'
+              gpu_model: 'Tesla T4'
+          - targets: ['140.44.4.83:9400']
+            labels:
+              node: 'snaisvr44hrvstr03'
+              gpu_model: 'Tesla T4'
+          - targets: ['140.44.4.84:9400']
+            labels:
+              node: 'snaisvr44hrvstr04'
+              gpu_model: 'Tesla T4'
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: prometheus
+  namespace: default
+  labels:
+    app.kubernetes.io/name: prometheus
+spec:
+  replicas: 1
+  strategy:
+    type: Recreate
+  selector:
+    matchLabels:
+      app.kubernetes.io/name: prometheus
+  template:
+    metadata:
+      labels:
+        app.kubernetes.io/name: prometheus
+    spec:
+      securityContext:
+        fsGroup: 65534
+        runAsUser: 65534
+        runAsNonRoot: true
+      containers:
+        - name: prometheus
+          image: prom/prometheus:v2.54.1
+          imagePullPolicy: IfNotPresent
+          args:
+            - "--config.file=/etc/prometheus/prometheus.yml"
+            - "--storage.tsdb.path=/prometheus"
+            - "--storage.tsdb.retention.time=90d"
+            - "--web.enable-lifecycle"
+            - "--web.console.libraries=/usr/share/prometheus/console_libraries"
+            - "--web.console.templates=/usr/share/prometheus/consoles"
+          ports:
+            - name: web
+              containerPort: 9090
+          resources:
+            requests:
+              cpu: "500m"
+              memory: "1Gi"
+            limits:
+              cpu: "2"
+              memory: "4Gi"
+          readinessProbe:
+            httpGet:
+              path: /-/ready
+              port: 9090
+            initialDelaySeconds: 10
+            periodSeconds: 5
+          livenessProbe:
+            httpGet:
+              path: /-/healthy
+              port: 9090
+            initialDelaySeconds: 30
+            periodSeconds: 15
+          volumeMounts:
+            - name: config
+              mountPath: /etc/prometheus
+            - name: storage
+              mountPath: /prometheus
+      volumes:
+        - name: config
+          configMap:
+            name: prometheus-config
+        - name: storage
+          persistentVolumeClaim:
+            claimName: prometheus-storage
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: prometheus
+  namespace: default
+  labels:
+    app.kubernetes.io/name: prometheus
+spec:
+  type: ClusterIP
+  selector:
+    app.kubernetes.io/name: prometheus
+  ports:
+    - name: web
+      port: 9090
+      targetPort: 9090
diff --git a/manifests/spacedock/ai-stack/ai-stack.yaml b/manifests/spacedock/ai-stack/ai-stack.yaml
index d0a580a..5a415ed 100644
--- a/manifests/spacedock/ai-stack/ai-stack.yaml
+++ b/manifests/spacedock/ai-stack/ai-stack.yaml
@@ -609,3 +609,53 @@ spec:
       port: 8002
       targetPort: 8002
       nodePort: 30803
+
+---
+# ==============================================================================
+# 4. NVIDIA DCGM EXPORTER (Tesla T4 Telemetry: Utilization, VRAM, Power, Temp)
+# ==============================================================================
+apiVersion: apps/v1
+kind: DaemonSet
+metadata:
+  name: dcgm-exporter
+  namespace: ai-stack
+  labels:
+    app.kubernetes.io/name: dcgm-exporter
+spec:
+  selector:
+    matchLabels:
+      app.kubernetes.io/name: dcgm-exporter
+  template:
+    metadata:
+      labels:
+        app.kubernetes.io/name: dcgm-exporter
+    spec:
+      hostNetwork: true
+      containers:
+      - name: dcgm-exporter
+        image: nvcr.io/nvidia/k8s/dcgm-exporter:3.3.8-3.6.0-ubuntu22.04
+        env:
+        - name: DCGM_EXPORTER_LISTEN
+          value: ":9400"
+        - name: DCGM_EXPORTER_KUBERNETES
+          value: "false"
+        - name: LD_LIBRARY_PATH
+          value: "/usr/local/nvidia/lib64"
+        securityContext:
+          privileged: true
+        ports:
+        - name: metrics
+          containerPort: 9400
+        volumeMounts:
+        - name: dev
+          mountPath: /dev
+        - name: nvidia-lib
+          mountPath: /usr/local/nvidia/lib64
+          readOnly: true
+      volumes:
+      - name: dev
+        hostPath:
+          path: /dev
+      - name: nvidia-lib
+        hostPath:
+          path: /usr/local/nvidia/lib64