Add custom metrics to FastAPI Server with prometheus_client | Python Application Monitoring

Monitoring the performance and health of a web application is crucial for ensuring its reliability and stability. In Python, one popular tool for application monitoring is prometheus_client, a Python client for Prometheus, a widely-used open-source monitoring and alerting toolkit.

FastAPI is a modern, high-performance web framework for building APIs with Python. In this article, we will explore how to add custom metrics to a FastAPI server using prometheus_client for application monitoring.

Setting up prometheus_client in FastAPI

First, we need to install the prometheus_client library in our Python environment using pip:

pip install prometheus_client

Next, we need to import the necessary modules from prometheus_client in our FastAPI application. We can then create custom metrics using the Counter, Gauge, or Summary classes provided by prometheus_client.

Creating custom metrics

Let’s say we want to track the number of requests to a specific endpoint in our FastAPI server. We can create a Counter metric to accomplish this:

from prometheus_client import Counter
from fastapi import FastAPI

app = FastAPI()

request_counter = Counter('http_requests_total', 'Total number of HTTP requests')

@app.get("/endpoint")
async def endpoint():
    request_counter.inc()
    return {"message": "Hello, World"}

In the code above, we create a Counter metric called http_requests_total with the description “Total number of HTTP requests”. We then use the inc() method to increment the counter every time the /endpoint route is accessed.

Exposing metrics endpoint

To make our custom metrics accessible to Prometheus, we need to expose an endpoint in our FastAPI server for scraping the metrics. We can do this by adding the following route to our FastAPI application:

from prometheus_client import generate_latest, CONTENT_TYPE_LATEST

@app.get("/metrics")
async def metrics():
    return Response(generate_latest(), media_type=CONTENT_TYPE_LATEST)

With the /metrics route added, we can now access our custom metrics using a Prometheus server for monitoring and visualization.

Conclusion

In this article, we’ve demonstrated how to add custom metrics to a FastAPI server using prometheus_client for application monitoring. By incorporating custom metrics into our FastAPI applications, we gain valuable insights into the performance and behavior of our web services, enabling us to make informed decisions and ensure the reliability of our applications.