How to filter out metrics for unhandled paths in Starlette Prometheus?

Pass `filter_unhandled_paths=True` when adding the PrometheusMiddleware to your Starlette app. This prevents accumulating data for routes not defined in your application, which is essential to avoid memory issues, as noted in the README.

Starlette Prometheus vs prometheus-client for Python?

Starlette Prometheus is tailored for Starlette, offering easier integration with automatic HTTP metrics. In contrast, prometheus-client is more general-purpose and requires manual setup for framework-specific metrics, making it better for non-Starlette projects.

Does Starlette Prometheus support custom metrics?

No, it focuses on HTTP request metrics like latency and counts. For custom metrics, you'll need to use the standard Prometheus Python client alongside it, which adds complexity and isn't documented in the README.

What metrics are exposed by Starlette Prometheus?

It exposes standard Prometheus metrics for HTTP requests, such as request duration, count, and status codes, via the `/metrics` endpoint. These help monitor application health but don't include application-level details without extra work.

Is Starlette Prometheus compatible with FastAPI?

Yes, since FastAPI is built on Starlette, you can use Starlette Prometheus with it by adding the middleware and metrics route similarly. However, ensure compatibility with your FastAPI version, as it depends on Starlette 0.12+.

How to set up Prometheus to scrape metrics from Starlette Prometheus?

After integrating the library, configure Prometheus to scrape the `/metrics` endpoint exposed by your Starlette app. This typically involves adding the app's URL to the Prometheus configuration file, but the README doesn't provide detailed guidance on Prometheus setup.

Open-Awesome

Starlette Prometheus

GPL-3.0Pythonv0.9.0

A Prometheus middleware and metrics endpoint for Starlette applications.

GitHub

376 stars33 forks0 contributors

What is Starlette Prometheus?

Starlette Prometheus is a Python library that integrates Prometheus monitoring into Starlette web applications. It provides middleware to automatically collect HTTP metrics and exposes them via a standard `/metrics` endpoint for Prometheus scraping. This solves the problem of monitoring the performance and health of asynchronous web services built with Starlette.

Target Audience

Developers building web applications with Starlette who need to monitor their services in production environments, particularly those using Prometheus for observability.

Value Proposition

It offers a simple, lightweight way to add comprehensive monitoring to Starlette apps without significant overhead, with features like optional path filtering to manage memory usage effectively.

Overview

Prometheus integration for Starlette.

Use Cases

Best For

Monitoring HTTP request latency and counts in Starlette applications
Exposing Prometheus metrics from ASGI-based web services
Adding observability to microservices built with Starlette
Tracking performance of API endpoints in production
Integrating Starlette apps into existing Prometheus monitoring stacks
Collecting custom metrics for asynchronous web frameworks

Not Ideal For

Projects using web frameworks other than Starlette, such as Django or Flask
Applications requiring advanced, application-specific metrics beyond HTTP request data
Teams already invested in a different monitoring system like Datadog or New Relic

Pros & Cons

Pros

Easy Integration

Setup is minimal, requiring only the addition of middleware and a route, as demonstrated in the README example with just a few lines of code.

Automatic HTTP Metrics

Collects essential performance data like response times and request counts automatically, eliminating the need for manual instrumentation in Starlette apps.

Memory Management Feature

Offers optional path filtering to prevent unbounded memory use by ignoring unhandled routes, a critical safeguard mentioned in the README for production environments.

Lightweight Design

Follows an unobtrusive philosophy, integrating monitoring without complicating the Starlette application structure, as stated in the project description.

Cons

Limited Metric Scope

Only provides basic HTTP metrics out of the box; custom or application-specific metrics require additional work with the Prometheus client, which isn't covered in the documentation.

Memory Footprint Risk

The README admits that without enabling path filtering, it can lead to unbounded memory use, posing a potential pitfall for developers who overlook this setting.

Framework Lock-in

Tightly coupled with Starlette, making it unsuitable for other ASGI frameworks or synchronous web frameworks without significant modifications or alternative libraries.

Frequently Asked Questions

Related Projects

Strawberry GraphQL

A GraphQL library for Python that leverages type annotations 🍓

Stars4,713

Forks660

Last commit15 hours ago

FastAPI Utilities

Reusable utilities for FastAPI

Stars2,308

Forks189

Last commit1 year ago

SlowApi

A rate limiter for Starlette and FastAPI

Stars2,056

Forks130

Last commit14 hours ago

FastAPI Cache

fastapi-cache is a tool to cache fastapi response and function result, with backends support redis and memcached.

Stars1,867

Forks206