IDE Metrics Component

Overview

The IDE Metrics component is a service in Gitpod that collects, processes, and exposes metrics and error reports from various IDE components. It provides a centralized way to gather performance data, usage statistics, and error information from IDE-related components such as the supervisor, VS Code extensions, and web workbenches.

Purpose

The primary purposes of the IDE Metrics component are:

• Collect metrics from IDE components
• Process and validate metric data
• Expose metrics in Prometheus format
• Handle error reporting from IDE components
• Provide a gRPC and REST API for metrics submission
• Support aggregated histogram metrics for efficient data collection
• Implement security measures through label allowlists
• Enable monitoring of IDE performance and usage

Architecture

The IDE Metrics component is built as a Go service with several key components:

1. Metrics Server: Core server that handles gRPC and HTTP requests
2. Metrics Registry: Manages Prometheus metrics
3. Error Reporter: Processes and forwards error reports
4. Label Validator: Validates metric labels against allowlists
5. Aggregated Histograms: Efficiently handles histogram data

The component exposes both gRPC and HTTP endpoints for metrics submission and provides a Prometheus-compatible endpoint for metrics scraping.

Key Files and Structure

• main.go: Entry point for the application
• cmd/root.go: Command-line interface setup
• cmd/run.go: Main server run command
• pkg/server/server.go: Core server implementation
• pkg/metrics/: Metrics handling utilities
• pkg/errorreporter/: Error reporting functionality
• config-example.json: Example configuration

Metrics Types

The IDE Metrics component supports several types of metrics:

Counter Metrics

Simple counters that can be incremented or added to, used for tracking occurrences of events.

Histogram Metrics

Record observations in buckets, used for measuring durations or sizes.

Aggregated Histogram Metrics

Efficiently handle pre-aggregated histogram data, reducing the number of individual metric submissions.

Error Reporting

The component includes an error reporting system that:

• Validates error reports against an allowlist of components
• Structures error data with contextual information
• Forwards errors to an error reporting service
• Includes workspace and user context with errors

Security Measures

The IDE Metrics component implements several security measures:

Label Allowlists

• Validates metric labels against configured allowlists
• Replaces invalid labels with default or "unknown" values
• Logs unexpected label names or values
• Prevents metric explosion from arbitrary labels

Component Allowlists

• Validates error reporting components against an allowlist
• Prevents unauthorized components from submitting errors
• Logs attempts from unexpected components

Configuration

The IDE Metrics component is configured through a JSON configuration file:

{
  "server": {
    "port": 9500,
    "counterMetrics": [
      {
        "name": "example_counter",
        "help": "Example counter metric",
        "labels": [
          {
            "name": "label1",
            "allowValues": ["value1", "value2"],
            "defaultValue": "unknown"
          }
        ]
      }
    ],
    "histogramMetrics": [...],
    "aggregatedHistogramMetrics": [...],
    "errorReporting": {
      "allowComponents": ["supervisor", "vscode-web"]
    }
  },
  "prometheus": {
    "addr": "localhost:9500"
  }
}

API Endpoints

The IDE Metrics component exposes several API endpoints:

• /metrics-api/: Base path for all API requests
• /metrics: Prometheus metrics endpoint

Dependencies

Internal Dependencies

• components/common-go: Common Go utilities
• components/ide-metrics-api: API definitions

External Dependencies

• Prometheus client for metrics
• gRPC for API communication
• gRPC-web for browser compatibility
• cmux for multiplexing connections

Integration Points

The IDE Metrics component integrates with:

1. Supervisor: Submits metrics and error reports
2. VS Code Extensions: Submit metrics and error reports
3. Web Workbench: Submits metrics and error reports
4. Prometheus: Scrapes exposed metrics
5. Error Reporting Service: Receives error reports

Common Usage Patterns

The IDE Metrics component is typically used to:

1. Track IDE performance metrics
2. Monitor feature usage
3. Collect error reports from IDE components
4. Generate dashboards for IDE health and usage
5. Identify performance bottlenecks

Related Components

• IDE Service: Configures and manages IDEs
• IDE Proxy: Proxies requests to IDE Metrics
• Supervisor: Submits metrics and error reports
• Dashboard: May display metrics data