Struct MetricsEndpoint 

pub struct MetricsEndpoint { /* private fields */ }

Lightweight HTTP server for exposing Prometheus metrics and health check endpoints.

MetricsEndpoint provides a simple, efficient HTTP server specifically designed for serving metrics to monitoring systems like Prometheus. It handles concurrent requests asynchronously and provides both metrics and health check endpoints.

Features

Prometheus Integration

• Standard Format: Serves metrics in Prometheus exposition format
• Content-Type: Proper content-type headers for Prometheus compatibility
• Fresh Data: Generates metrics fresh on each request
• Error Handling: Graceful error responses for metrics generation failures

Health Monitoring

• Health Endpoint: Simple health check endpoint for load balancers
• Fast Response: Low-latency health check responses
• Status Codes: Proper HTTP status codes for health states
• Monitoring Integration: Compatible with various monitoring systems

Performance Optimized

• Async Processing: Non-blocking request handling
• Concurrent Connections: Handles multiple simultaneous requests
• Low Overhead: Minimal resource usage
• Efficient Parsing: Simple HTTP request parsing

Usage Examples

Basic Server Setup

Integration with Tokio Runtime

Configuration

The server is configured through the ConfigService:

• Port number for the HTTP server
• Bind address (currently localhost only)
• Request handling timeouts

Error Handling

The server handles various error conditions gracefully:

• Metrics Generation Errors: Returns HTTP 500 with error details
• Connection Errors: Logs errors and continues serving other requests
• Invalid Requests: Returns HTTP 404 for unknown endpoints
• Resource Exhaustion: Graceful degradation under high load

Thread Safety

The endpoint is thread-safe and can be used concurrently:

• Immutable after construction
• Shared metrics service through Arc
• Safe concurrent request handling

Implementations

impl MetricsEndpoint

pub fn new(metrics_service: Arc<MetricsService>) -> Self

Creates a new metrics endpoint server with the provided metrics // service.

Arguments

• metrics_service - Arc-wrapped metrics service for generating metrics data

Returns

A new MetricsEndpoint instance ready to start serving requests.

Examples

pub async fn start(&self) -> Result<(), PipelineError>

Starts the metrics endpoint HTTP server.

This method starts the HTTP server and begins accepting connections. It runs indefinitely, handling incoming requests concurrently. The server will bind to the configured port and serve metrics and health check endpoints.

Returns

• Ok(()) - Never returns normally (runs indefinitely)
• Err(PipelineError) - If server fails to start or bind to port

Endpoints Served

• GET /metrics - Prometheus metrics in text format
• GET /health - Simple health check returning “OK”
• Other paths return 404 Not Found

Examples

Error Handling

The server handles individual request errors gracefully:

• Connection errors are logged but don’t stop the server
• Metrics generation errors return HTTP 500
• Invalid requests return HTTP 404

Performance

• Handles requests concurrently using Tokio tasks
• Low memory overhead per connection
• Efficient request parsing and response generation