fastapi-insights 0.1.2

fastapi-insights is a FastAPI extension that tracks request and system metrics and provides a built-in dashboard.

fastapi-insights

Introduction

fastapi-insights is a FastAPI extension for application performance monitoring. It tracks request and system metrics using lightweight middleware. Metrics can be stored in in-memory, SQLite, or Redis backends and visualized in a built-in dashboard UI.

Who is it for?

• Developers who want a metrics dashboard without running a full Prometheus + Grafana stack.
• Indie devs or small teams running single-instance FastAPI backends who just need lightweight insights.

Features

• 🚀 Zero-config FastAPI middleware
• 🗄 Multiple storage backends: in-memory, SQLite, Redis
• 💻 Built-in dashboard UI with charts
• ⚡ Lightweight & async-first design
• 🔌 Configurable retention, bucket sizes, and cleanup

Installation

> pip install fastapi-insights

Optional dependencies:

> pip install "fastapi-insights[redis]"

Supported Python versions: 3.10+

Development

Install dev dependencies and set up pre-commit hooks:

> uv sync --dev
> uv run pre-commit install

Run the checks manually at any time with:

> uv run pre-commit run --all-files

Quick Start

Check the examples folder for more.

from fastapi import FastAPI
from fastapi_insights import FastAPIInsights, Config
from fastapi_insights.backends.in_memory import InMemoryMetricsStore

app = FastAPI()

FastAPIInsights.init(
    app,
    InMemoryMetricsStore(),
    config=Config(),
)

@app.get("/")
def index():
    return "ok"

Visit /metrics to view the UI.

Backends

In-Memory (default)

from fastapi_insights.backends.in_memory import InMemoryMetricsStore
store = InMemoryMetricsStore()

Redis

from fastapi_insights.backends.redis import AsyncRedisMetricsStore, RedisMetricsStore

# sync
import redis

redis_client = redis.Redis(host="localhost", port=6379, db=0)
store = RedisMetricsStore(redis_client)

# async
import redis.asyncio as async_redis

async_redis_client = async_redis.Redis(host="localhost", port=6379, db=0)
async_store = AsyncRedisMetricsStore(async_redis_client)

SQLite

from fastapi_insights.backends.sqlite import SQLiteMetricsStore
store = SQLiteMetricsStore("metrics.db")

The SQLite backend uses the standard library sqlite3 module, so no extra dependency is required.

Configuration

You can customize the behavior by adjusting the Config options at initialization.

Example

from fastapi_insights import Config

config = Config(
    ignored_routes=["/health", "/internal/*"],
    exclude_library_metrics=True,
    ui_config_route="/config-b887e852-bd12-41f2-b057-1bd31eb5443e",
    enable_dashboard_ui=True,
    custom_path="/metrics",
    include_in_openapi=False,
    ui_title="FastAPI Metrics"
)

Available Options

Option	Type	Default	Description
ignored_routes	list[str]	[]	Routes to be ignored by the metrics middleware. Use exact matches like /route, or wildcard matches like /route/* for prefixes.
exclude_library_metrics	bool	True	If True, excludes internal fastapi-insights endpoints from being tracked.
ui_config_route	str	"/config-b887e852-bd12-41f2-b057-1bd31eb5443e"	Internal route used to serve the dashboard’s configuration JSON. Should generally be left unchanged.
enable_dashboard_ui	bool	True	Enables or disables the built-in dashboard UI. When disabled, only raw metrics endpoints remain active.
custom_path	str	"/metrics"	The URL path under which metrics are exposed (e.g., /metrics or /stats).
include_in_openapi	bool	False	Whether to include the metrics endpoints in your FastAPI OpenAPI schema (Swagger UI).
ui_title	str	"FastAPI Metrics"	The title displayed in the dashboard UI.

Development

git clone https://github.com/HamzaDLM/fastapi-insights
cd fastapi-insights
uv sync
uv run pytest

Limitations

• Best suited for single-instance FastAPI deployments or small teams that want an embedded dashboard.
• Not a replacement for a full metrics pipeline such as Prometheus + Grafana in larger distributed systems.
• The in-memory backend is process-local and not durable across restarts.

License

This project is licensed under the Apache-2.0 License.