@senad-d/observme 0.1.0

package/dashboards/observme-cost.json

@@ -0,0 +1,1254 @@
+{
+  "__requires": [
+    {
+      "id": "grafana",
+      "name": "Grafana",
+      "type": "grafana",
+      "version": "10.0.0"
+    },
+    {
+      "id": "prometheus",
+      "name": "Prometheus",
+      "type": "datasource",
+      "version": "1.0.0"
+    },
+    {
+      "id": "row",
+      "name": "Row",
+      "type": "panel",
+      "version": ""
+    },
+    {
+      "id": "timeseries",
+      "name": "Time series",
+      "type": "panel",
+      "version": ""
+    },
+    {
+      "id": "stat",
+      "name": "Stat",
+      "type": "panel",
+      "version": ""
+    },
+    {
+      "id": "bargauge",
+      "name": "Bar gauge",
+      "type": "panel",
+      "version": ""
+    },
+    {
+      "id": "text",
+      "name": "Text",
+      "type": "panel",
+      "version": ""
+    }
+  ],
+  "annotations": {
+    "list": [
+      {
+        "datasource": {
+          "type": "prometheus",
+          "uid": "prometheus"
+        },
+        "enable": true,
+        "hide": false,
+        "iconColor": "rgba(31, 120, 193, 1)",
+        "name": "Model changes",
+        "expr": "sum(increase(observme_model_changes_total[$__interval])) by (provider, model) > 0",
+        "titleFormat": "Model change",
+        "textFormat": "{{provider}} / {{model}}"
+      },
+      {
+        "datasource": {
+          "type": "prometheus",
+          "uid": "prometheus"
+        },
+        "enable": true,
+        "hide": false,
+        "iconColor": "rgba(245, 183, 0, 1)",
+        "name": "Thinking level changes",
+        "expr": "sum(increase(observme_thinking_level_changes_total[$__interval])) by (environment) > 0",
+        "titleFormat": "Thinking level change",
+        "textFormat": "{{environment}}"
+      }
+    ]
+  },
+  "editable": true,
+  "fiscalYearStartMonth": 0,
+  "graphTooltip": 1,
+  "id": null,
+  "links": [
+    {
+      "icon": "external link",
+      "includeVars": true,
+      "keepTime": true,
+      "targetBlank": false,
+      "title": "Overview",
+      "tooltip": "Open the overview with the current time range.",
+      "type": "link",
+      "url": "/d/observme-overview/observme-overview?${__url_time_range}"
+    },
+    {
+      "icon": "external link",
+      "includeVars": true,
+      "keepTime": true,
+      "targetBlank": false,
+      "title": "Models",
+      "tooltip": "Open model traffic, latency, stop-reason, and cost-efficiency panels.",
+      "type": "link",
+      "url": "/d/observme-models/observme-models?${__url_time_range}"
+    },
+    {
+      "icon": "external link",
+      "includeVars": true,
+      "keepTime": true,
+      "targetBlank": false,
+      "title": "LLM lifecycle logs",
+      "tooltip": "Open LLM token and non-content lifecycle logs.",
+      "type": "link",
+      "url": "/d/observme-logs-llm/observme-logs-llm?${__url_time_range}"
+    },
+    {
+      "icon": "external link",
+      "includeVars": true,
+      "keepTime": true,
+      "targetBlank": false,
+      "title": "LLM Conversations",
+      "tooltip": "Open optional captured LLM content with privacy guidance.",
+      "type": "link",
+      "url": "/d/observme-llm-conversations/observme-llm-conversations?${__url_time_range}"
+    },
+    {
+      "icon": "external link",
+      "includeVars": true,
+      "keepTime": true,
+      "targetBlank": false,
+      "title": "Errors",
+      "tooltip": "Open LLM failure ratios, failure logs, and trace links.",
+      "type": "link",
+      "url": "/d/observme-errors/observme-errors?${__url_time_range}"
+    }
+  ],
+  "liveNow": false,
+  "panels": [
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 0
+      },
+      "id": 100,
+      "panels": [],
+      "title": "Spend overview",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Total provider-reported LLM spend in the selected dashboard range. Missing cost series render as zero. The $40/$50 colors are review references; the $50 alert threshold applies specifically to a one-hour window.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "yellow",
+                "value": 40
+              },
+              {
+                "color": "red",
+                "value": 50
+              }
+            ]
+          },
+          "unit": "currencyUSD"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 0,
+        "y": 1
+      },
+      "id": 1,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(increase(observme_llm_cost_usd_total[$__range])) or vector(0)",
+          "instant": true,
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Total selected-range cost",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Average hourly LLM cost estimated from selected-range spend. Yellow/red align with 80%/100% of the default $50-in-1h cost-spike alert threshold; tune for your deployment.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "yellow",
+                "value": 40
+              },
+              {
+                "color": "red",
+                "value": 50
+              }
+            ]
+          },
+          "unit": "currencyUSD"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 6,
+        "y": 1
+      },
+      "id": 2,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(sum(increase(observme_llm_cost_usd_total[$__range])) or vector(0)) / ($__range_s / 3600)",
+          "instant": true,
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Cost burn rate per hour",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Selected-range spend divided by the default $50/hour budget scaled to the dashboard range length. This is an average budget-envelope comparison, not the state of the rolling one-hour cost-spike alert.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "yellow",
+                "value": 0.8
+              },
+              {
+                "color": "red",
+                "value": 1
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 12,
+        "y": 1
+      },
+      "id": 3,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(sum(increase(observme_llm_cost_usd_total[$__range])) or vector(0)) / (50 * ($__range_s / 3600))",
+          "instant": true,
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Selected-range budget usage",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Selected-range average burn extrapolated to 24 hours. Yellow/red align with 80%/100% of the default $1,200 daily equivalent; short or unrepresentative ranges can be noisy.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              },
+              {
+                "color": "yellow",
+                "value": 960
+              },
+              {
+                "color": "red",
+                "value": 1200
+              }
+            ]
+          },
+          "unit": "currencyUSD"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 6,
+        "x": 18,
+        "y": 1
+      },
+      "id": 4,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(sum(increase(observme_llm_cost_usd_total[$__range])) or vector(0)) * 86400 / $__range_s",
+          "instant": true,
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Projected 24h cost",
+      "type": "stat"
+    },
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 6
+      },
+      "id": 101,
+      "panels": [],
+      "title": "Cost attribution",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Top provider cost totals over the selected range. Missing cost series render as zero; open Models for traffic and cost-efficiency context.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "currencyUSD"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 8,
+        "x": 0,
+        "y": 7
+      },
+      "id": 5,
+      "options": {
+        "displayMode": "gradient",
+        "maxVizHeight": 300,
+        "minVizHeight": 16,
+        "minVizWidth": 8,
+        "namePlacement": "auto",
+        "orientation": "horizontal",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showUnfilled": true,
+        "sizing": "auto",
+        "valueMode": "color"
+      },
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(topk(10, sum by (provider) (increase(observme_llm_cost_usd_total[$__range]))) or on() vector(0))",
+          "instant": true,
+          "legendFormat": "{{provider}}",
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Cost by provider",
+      "type": "bargauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Top provider/model cost totals over the selected range. Keeping provider in the grouping prevents identically named models from different providers being merged.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "currencyUSD"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 8,
+        "x": 8,
+        "y": 7
+      },
+      "id": 6,
+      "options": {
+        "displayMode": "gradient",
+        "maxVizHeight": 300,
+        "minVizHeight": 16,
+        "minVizWidth": 8,
+        "namePlacement": "auto",
+        "orientation": "horizontal",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showUnfilled": true,
+        "sizing": "auto",
+        "valueMode": "color"
+      },
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(topk(10, sum by (provider, model) (increase(observme_llm_cost_usd_total[$__range]))) or on() vector(0))",
+          "instant": true,
+          "legendFormat": "{{provider}} / {{model}}",
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Cost by provider/model",
+      "type": "bargauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Selected-range provider-reported LLM cost grouped by deployment environment. Missing cost series render as zero.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "currencyUSD"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 8,
+        "x": 16,
+        "y": 7
+      },
+      "id": 7,
+      "options": {
+        "displayMode": "gradient",
+        "maxVizHeight": 300,
+        "minVizHeight": 16,
+        "minVizWidth": 8,
+        "namePlacement": "auto",
+        "orientation": "horizontal",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showUnfilled": true,
+        "sizing": "auto",
+        "valueMode": "color"
+      },
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(sum by (environment) (increase(observme_llm_cost_usd_total[$__range])) or on() vector(0))",
+          "instant": true,
+          "legendFormat": "{{environment}}",
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Cost by environment",
+      "type": "bargauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Top provider/model shares of total selected-range LLM spend. This adds explicit percent-of-total context beside the absolute attribution panels; missing cost series render as zero.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 7,
+        "w": 24,
+        "x": 0,
+        "y": 15
+      },
+      "id": 14,
+      "links": [
+        {
+          "targetBlank": false,
+          "title": "Open Models",
+          "url": "/d/observme-models/observme-models?${__url_time_range}"
+        }
+      ],
+      "options": {
+        "displayMode": "gradient",
+        "maxVizHeight": 300,
+        "minVizHeight": 16,
+        "minVizWidth": 8,
+        "namePlacement": "auto",
+        "orientation": "horizontal",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showUnfilled": true,
+        "sizing": "auto",
+        "valueMode": "color"
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(topk(10, sum by (provider, model) (increase(observme_llm_cost_usd_total[$__range])) / scalar(clamp_min((sum(increase(observme_llm_cost_usd_total[$__range])) or vector(0)), 1e-9))) or on() vector(0))",
+          "instant": true,
+          "legendFormat": "{{provider}} / {{model}}",
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Cost share by provider/model",
+      "type": "bargauge"
+    },
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 22
+      },
+      "id": 102,
+      "panels": [],
+      "title": "Cost trend",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "LLM cost added across all providers and models for each Prometheus-safe rolling interval. $__rate_interval keeps the window large enough for counter samples; small costs display with six decimal places.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "decimals": 6,
+          "mappings": [],
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "currencyUSD"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 24,
+        "x": 0,
+        "y": 23
+      },
+      "id": 8,
+      "links": [
+        {
+          "targetBlank": false,
+          "title": "Open Models",
+          "url": "/d/observme-models/observme-models?${__url_time_range}"
+        },
+        {
+          "targetBlank": false,
+          "title": "Open LLM lifecycle logs",
+          "url": "/d/observme-logs-llm/observme-logs-llm?${__url_time_range}"
+        }
+      ],
+      "options": {
+        "legend": {
+          "calcs": [
+            "sum",
+            "max"
+          ],
+          "displayMode": "table",
+          "placement": "bottom",
+          "showLegend": true
+        },
+        "tooltip": {
+          "mode": "multi",
+          "sort": "desc"
+        }
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(increase(observme_llm_cost_usd_total[$__rate_interval])) or vector(0)",
+          "legendFormat": "all providers / models",
+          "range": true,
+          "refId": "A"
+        }
+      ],
+      "title": "Cost over time",
+      "type": "timeseries"
+    },
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 31
+      },
+      "id": 103,
+      "panels": [],
+      "title": "Tokens and cache",
+      "type": "row"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Highest cumulative token value observed per exporter series in the selected range, summed globally by type. This preserves short-lived Pi session exports that counter increases can miss. Provider/model detail stays in Models and LLM lifecycle logs.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "short"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 0,
+        "y": 32
+      },
+      "id": 9,
+      "options": {
+        "displayMode": "gradient",
+        "maxVizHeight": 300,
+        "minVizHeight": 16,
+        "minVizWidth": 8,
+        "namePlacement": "auto",
+        "orientation": "horizontal",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showUnfilled": true,
+        "sizing": "auto",
+        "valueMode": "color"
+      },
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(max_over_time(observme_llm_input_tokens_total[$__range])) or vector(0)",
+          "instant": true,
+          "legendFormat": "input",
+          "range": false,
+          "refId": "A"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(max_over_time(observme_llm_output_tokens_total[$__range])) or vector(0)",
+          "instant": true,
+          "legendFormat": "output",
+          "range": false,
+          "refId": "B"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(max_over_time(observme_llm_total_tokens_total[$__range])) or vector(0)",
+          "instant": true,
+          "legendFormat": "total",
+          "range": false,
+          "refId": "C"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(max_over_time(observme_llm_reasoning_tokens_total[$__range])) or vector(0)",
+          "instant": true,
+          "legendFormat": "reasoning",
+          "range": false,
+          "refId": "D"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(max_over_time(observme_llm_cache_read_tokens_total[$__range])) or vector(0)",
+          "instant": true,
+          "legendFormat": "cache read",
+          "range": false,
+          "refId": "E"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(max_over_time(observme_llm_cache_write_tokens_total[$__range])) or vector(0)",
+          "instant": true,
+          "legendFormat": "cache write",
+          "range": false,
+          "refId": "F"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(max_over_time(observme_llm_cache_write_1h_tokens_total[$__range])) or vector(0)",
+          "instant": true,
+          "legendFormat": "cache write 1h",
+          "range": false,
+          "refId": "G"
+        }
+      ],
+      "title": "Token totals by type",
+      "type": "bargauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Top provider/model cache-read, cache-write, and one-hour cache-write token totals over the selected range. Missing cache metrics render as zero.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "green",
+                "value": null
+              }
+            ]
+          },
+          "unit": "short"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 8,
+        "w": 12,
+        "x": 12,
+        "y": 32
+      },
+      "id": 10,
+      "options": {
+        "displayMode": "gradient",
+        "maxVizHeight": 300,
+        "minVizHeight": 16,
+        "minVizWidth": 8,
+        "namePlacement": "auto",
+        "orientation": "horizontal",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showUnfilled": true,
+        "sizing": "auto",
+        "valueMode": "color"
+      },
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(topk(10, sum by (provider, model) (increase(observme_llm_cache_read_tokens_total[$__range]))) or on() vector(0))",
+          "instant": true,
+          "legendFormat": "cache read {{provider}} / {{model}}",
+          "range": false,
+          "refId": "A"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(topk(10, sum by (provider, model) (increase(observme_llm_cache_write_tokens_total[$__range]))) or on() vector(0))",
+          "instant": true,
+          "legendFormat": "cache write {{provider}} / {{model}}",
+          "range": false,
+          "refId": "B"
+        },
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "(topk(10, sum by (provider, model) (increase(observme_llm_cache_write_1h_tokens_total[$__range]))) or on() vector(0))",
+          "instant": true,
+          "legendFormat": "cache write 1h {{provider}} / {{model}}",
+          "range": false,
+          "refId": "C"
+        }
+      ],
+      "title": "Cache read/write token usage",
+      "type": "bargauge"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Estimated cache-read share of reusable prompt tokens: cache reads divided by input plus cache-read tokens. This is a proxy, not a provider hit-count or health metric; zero can mean idle workload or unsupported cache accounting.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "palette-classic"
+          },
+          "mappings": [],
+          "max": 1,
+          "min": 0,
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "blue",
+                "value": null
+              }
+            ]
+          },
+          "unit": "percentunit"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 12,
+        "x": 0,
+        "y": 40
+      },
+      "id": 11,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "clamp_max((sum(increase(observme_llm_cache_read_tokens_total[$__range])) or vector(0)) / clamp_min(((sum(increase(observme_llm_cache_read_tokens_total[$__range])) or vector(0)) + (sum(increase(observme_llm_input_tokens_total[$__range])) or vector(0))), 1), 1)",
+          "instant": true,
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Cache read ratio",
+      "type": "stat"
+    },
+    {
+      "datasource": {
+        "type": "prometheus",
+        "uid": "prometheus"
+      },
+      "description": "Provider-reported cache-read tokens over the selected range. Treat this as an estimated token saving only where cache-read accounting is supported; missing series render as zero.",
+      "fieldConfig": {
+        "defaults": {
+          "color": {
+            "mode": "thresholds"
+          },
+          "mappings": [],
+          "thresholds": {
+            "mode": "absolute",
+            "steps": [
+              {
+                "color": "gray",
+                "value": null
+              },
+              {
+                "color": "green",
+                "value": 1
+              }
+            ]
+          },
+          "unit": "short"
+        },
+        "overrides": []
+      },
+      "gridPos": {
+        "h": 5,
+        "w": 12,
+        "x": 12,
+        "y": 40
+      },
+      "id": 12,
+      "options": {
+        "colorMode": "value",
+        "graphMode": "area",
+        "justifyMode": "auto",
+        "orientation": "auto",
+        "reduceOptions": {
+          "calcs": [
+            "lastNotNull"
+          ],
+          "fields": "",
+          "values": false
+        },
+        "showPercentChange": false,
+        "textMode": "auto",
+        "wideLayout": true
+      },
+      "pluginVersion": "11.1.0",
+      "targets": [
+        {
+          "datasource": {
+            "type": "prometheus",
+            "uid": "prometheus"
+          },
+          "editorMode": "code",
+          "expr": "sum(increase(observme_llm_cache_read_tokens_total[$__range])) or vector(0)",
+          "instant": true,
+          "range": false,
+          "refId": "A"
+        }
+      ],
+      "title": "Estimated cache savings (tokens)",
+      "type": "stat"
+    },
+    {
+      "collapsed": false,
+      "gridPos": {
+        "h": 1,
+        "w": 24,
+        "x": 0,
+        "y": 45
+      },
+      "id": 104,
+      "panels": [],
+      "title": "Budget guidance",
+      "type": "row"
+    },
+    {
+      "description": "Visible interpretation guidance for the dashboard's default budget references and forecasts.",
+      "gridPos": {
+        "h": 5,
+        "w": 24,
+        "x": 0,
+        "y": 46
+      },
+      "id": 13,
+      "options": {
+        "code": {
+          "language": "markdown",
+          "showLineNumbers": false,
+          "showMiniMap": false
+        },
+        "content": "### Default budget references\n\n- **ObservMeCostSpike** alerts when LLM spend exceeds **$50 in 1 hour** for 15 minutes. Tune this budget-dependent threshold for your deployment.\n- **Selected-range budget usage** scales that $50/hour reference to the selected time range; 100% means spend matched the average envelope. It does not detect a short spike hidden inside a longer range.\n- **Projected 24h cost** extrapolates the average burn rate from the selected range. Very short, idle, or unrepresentative ranges can produce misleading forecasts.\n- Cost metrics use provider-reported usage. Missing cost series render as zero and do not prove that a provider supplied complete pricing data.",
+        "mode": "markdown"
+      },
+      "pluginVersion": "11.1.0",
+      "title": "How to interpret budget and forecast panels",
+      "type": "text"
+    }
+  ],
+  "refresh": "30s",
+  "schemaVersion": 39,
+  "tags": [
+    "observme",
+    "pi",
+    "llm",
+    "cost",
+    "tokens"
+  ],
+  "time": {
+    "from": "now-6h",
+    "to": "now"
+  },
+  "timezone": "browser",
+  "title": "ObservMe Cost",
+  "uid": "observme-cost",
+  "version": 8,
+  "weekStart": ""
+}