Add local AI metrics dashboard for self-hosted usage analytics

package.json

@@ -1992,6 +1992,12 @@
 				"icon": "$(feedback)",
 				"category": "Chat"
 			},
+			{
+				"command": "github.copilot.viewMetrics",
+				"title": "View AI Metrics Dashboard",
+				"icon": "$(graph)",
+				"category": "Copilot"
+			},
 			{
 				"command": "github.copilot.debug.workbenchState",
 				"title": "%github.copilot.command.logWorkbenchState%",
@@ -2439,6 +2445,18 @@
 						"type": "string",
 						"default": "",
 						"markdownDescription": "The currently selected completion model ID. To select from a list of available models, use the __\"Change Completions Model\"__ command or open the model picker (from the Copilot menu in the VS Code title bar, select __\"Configure Code Completions\"__ then __\"Change Completions Model\"__. The value must be a valid model ID. An empty value indicates that the default model will be used."
+					},
+					"github.copilot.metrics.enabled": {
+						"type": "boolean",
+						"default": false,
+						"markdownDescription": "Enable local AI metrics collection and dashboard. When enabled, Copilot usage metrics are stored locally for personal insights. This is separate from GitHub telemetry controls."
+					},
+					"github.copilot.metrics.retentionDays": {
+						"type": "number",
+						"default": 90,
+						"minimum": 7,
+						"maximum": 365,
+						"markdownDescription": "Number of days to retain local AI metrics data. Data older than this will be automatically pruned. Default is 90 days."
 					}
 				}
 			},

src/extension/aiMetrics/README.md

@@ -0,0 +1,136 @@
+# AI Metrics Dashboard
+
+A local developer dashboard for visualizing personal GitHub Copilot usage metrics. All data is stored locally and never leaves your machine.
+
+## Features
+
+- **Token Usage Tracking**: Monitor total tokens consumed, tokens by model, and cache efficiency
+- **Model Distribution**: See which AI models you use most frequently
+- **Code Acceptance Metrics**: Track acceptance rates for NES and completions
+- **Feature Usage**: Monitor chat messages, NES opportunities, and completions
+- **Performance Metrics**: View average and P95 latencies (TTFT, fetch time)
+
+## Configuration
+
+### Enable Metrics Collection
+
+Add to your VS Code settings:
+
+```json
+{
+  "github.copilot.metrics.enabled": true,
+  "github.copilot.metrics.retentionDays": 90
+}
+```
+
+### Settings
+
+- `github.copilot.metrics.enabled` (boolean, default: `false`)
+  - Enable local AI metrics collection
+  - Data is stored in VS Code's global state
+  - No data is sent to external services
+
+- `github.copilot.metrics.retentionDays` (number, default: `90`, min: `7`, max: `365`)
+  - Number of days to retain metrics data
+  - Old data is automatically pruned on extension activation
+  - Default is 90 days
+
+## Usage
+
+1. **Enable metrics collection** in settings (see above)
+
+2. **Use Copilot normally**:
+   - Chat with Copilot
+   - Accept/reject NES suggestions
+   - Use code completions
+   - All relevant telemetry events are captured
+
+3. **View the dashboard**:
+   - Open Command Palette (`Ctrl+Shift+P` or `Cmd+Shift+P`)
+   - Type: "View AI Metrics Dashboard"
+   - Or run command: `github.copilot.viewMetrics`
+
+4. **Select time range**:
+   - Today
+   - Last 7 Days (default)
+   - Last 30 Days
+   - All Time
+
+5. **Refresh metrics**:
+   - Click the "Refresh" button
+   - Metrics are computed on-demand (no background processing)
+
+## Dashboard Sections
+
+### Overview Cards
+
+- **Total Tokens**: Total tokens consumed with cache ratio
+- **Acceptance Rate**: NES suggestion acceptance percentage
+- **Top Model**: Most frequently used AI model
+- **Total Events**: Number of metric events collected
+
+### Detailed Charts
+
+- **Tokens by Model**: Breakdown of token usage across different models
+- **Tokens by Feature**: Token usage by Copilot feature (chat, NES, completions)
+- **Feature Usage**: Counts for chat messages, NES opportunities, and completions
+- **Performance Metrics**: Average and P95 latencies
+
+## Data Storage
+
+- **Location**: VS Code global state
+- **Schema**: `aiMetrics.events.<YYYY-MM-DD>[]`
+- **Format**: Events grouped by day for efficient storage
+- **Privacy**: All data stays local, never transmitted
+
+## Architecture
+
+### Components
+
+- **Domain Model** (`src/extension/aiMetrics/common/metrics.ts`)
+  - Metric type definitions
+  - Data structures for aggregated metrics
+
+- **Storage Service** (`src/extension/aiMetrics/node/aiMetricsStorageService.ts`)
+  - Stores events in VS Code global state
+  - Provides query and aggregation methods
+  - Handles data pruning
+
+- **Telemetry Collector** (`src/extension/aiMetrics/common/aiMetricsCollector.ts`)
+  - Intercepts telemetry events
+  - Extracts metric-relevant data
+  - Forwards to storage service
+
+- **Dashboard Panel** (`src/extension/aiMetrics/vscode-node/aiMetricsDashboardPanel.ts`)
+  - Webview-based UI
+  - On-demand metric computation
+  - Theme-aware styling
+
+### Event Types
+
+- **TokenUsage**: Tracks token consumption
+- **ModelUsage**: Records model usage
+- **CodeAcceptance**: Captures acceptance/rejection events
+- **FeatureUsage**: Monitors feature usage
+- **Performance**: Collects latency metrics
+
+## Development
+
+### Running Tests
+
+```bash
+npm run test:unit -- src/extension/aiMetrics/test
+```
+
+### Building
+
+```bash
+npm run compile
+```
+
+## Notes
+
+- Metrics collection has minimal performance impact
+- Dashboard only computes metrics when manually refreshed
+- No background processing or scheduled tasks
+- All data is local and under your control