FastAPI and Uvicorn is running synchronously and very slow
  • 31-May-2023
Lightrun Team
Author Lightrun Team
FastAPI and Uvicorn is running synchronously and very slow

FastAPI and Uvicorn is running synchronously and very slow

Lightrun Team
Lightrun Team

Explanation of the problem

The issue at hand involves testing file uploads and asynchronous requests in FastAPI. When performing multiple requests with clients in parallel or serially, FastAPI processes each upload sequentially, resulting in slow performance. The API is being executed using uvicorn and gunicorn with a single worker, but the execution time remains the same regardless of the chosen approach.

The client sends four files, each approximately 20MB in size, to a FastAPI endpoint either in parallel or serially. However, FastAPI processes the files one at a time, causing significant delays. In contrast, when performing the same uploads with an aiohttp endpoint, the files are stored much faster, taking around 0.6 seconds for parallel requests (using multiprocessing) and 0.8 seconds for serial requests on average.

The FastAPI server code provided demonstrates the handling of file uploads. Upon receiving an upload, it creates a unique file path in the storage directory and uses aiofiles to asynchronously write the content of the file. However, despite implementing asynchronous handling, the performance remains slow compared to the aiohttp implementation. The client code utilizes the requests library to make parallel and serial requests, measuring the elapsed time for the upload process.


Troubleshooting with the Lightrun Developer Observability Platform


Getting a sense of what’s actually happening inside a live application is a frustrating experience, one that relies mostly on querying and observing whatever logs were written during development.
Lightrun is a Developer Observability Platform, allowing developers to add telemetry to live applications in real-time, on-demand, and right from the IDE.

  • Instantly add logs to, set metrics in, and take snapshots of live applications
  • Insights delivered straight to your IDE or CLI
  • Works where you do: dev, QA, staging, CI/CD, and production

Start for free today

Problem solution FastAPI and Uvicorn is running synchronously and very slow

One approach to improving the performance of handling multipart/form-data requests in FastAPI is by modifying the default spooling behavior of the UploadFile class. By setting the spool_max_size attribute to 0, the UploadFile will keep the data in memory using a SpooledTemporaryFile, rather than rolling it over onto disk. This adjustment can eliminate the file I/O trip and potentially provide performance similar to that of frameworks like Sanic and uvicorn. By reducing the need for file operations, the processing of multipart data can become more efficient.

Another solution to enhance the performance of multipart/form-data parsing in FastAPI is to leverage the streaming_form_data library. This library, implemented in Cython, offers faster multipart parsing compared to the default implementation in starlette. The author of the library mentions that pure Python byte-wise parsing is slower, which impacts the performance of multipart handling. By utilizing streaming_form_data, developers can take advantage of the optimized parsing capabilities and achieve faster processing of multipart/form-data requests.

In summary, there are multiple techniques available to improve the performance of handling multipart/form-data requests in FastAPI. Modifying the spooling behavior of the UploadFile class can reduce file I/O and potentially enhance performance. Additionally, utilizing external libraries like streaming_form_data, implemented in Cython, can offer faster multipart parsing by leveraging optimized byte-wise parsing techniques. Developers can choose the approach that best fits their specific requirements and take advantage of these optimizations to achieve efficient and high-performing multipart handling in FastAPI.

Problems with fastapi


Problem 1: CORS (Cross-Origin Resource Sharing) Issues

One common problem encountered with FastAPI is dealing with CORS issues. When making cross-origin requests from a web browser, the browser enforces security policies that restrict access to resources on different domains. FastAPI applications hosted on one domain may encounter CORS errors when attempting to make requests to another domain.

To solve this issue, FastAPI provides a middleware called fastapi.middleware.cors. By adding this middleware to your application, you can configure the necessary CORS headers to allow cross-origin requests. Here’s an example of how to use the CORS middleware in FastAPI:


from fastapi import FastAPI
from fastapi.middleware.cors import CORSMiddleware

app = FastAPI()

# Configure CORS settings
origins = [



In the code snippet above, we create a list of allowed origins that are allowed to make cross-origin requests. We then add the CORSMiddleware to the FastAPI application using app.add_middleware(), passing in the necessary configuration options. This ensures that the necessary CORS headers are included in the server responses, allowing cross-origin requests to be made successfully.

Problem 2: Request Validation and Serialization

Another common problem in FastAPI is request validation and serialization. FastAPI provides powerful request validation capabilities out of the box, allowing you to define request models and automatically validate and deserialize incoming requests. However, when dealing with complex data structures or nested models, ensuring proper validation and serialization can become challenging.

To address this issue, you can leverage Pydantic, which is integrated with FastAPI and provides robust data validation and serialization features. By creating Pydantic models for your request bodies, you can define the expected structure, types, and constraints. Here’s an example:


from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI()

class Item(BaseModel):
    name: str
    price: float"/items/")
async def create_item(item: Item):
    # Access validated and deserialized item properties
    item_name =
    item_price = item.price
    # ...


In the code snippet above, we define a Item Pydantic model that represents the structure of the request body. By specifying the expected types and constraints, FastAPI will automatically validate and deserialize the incoming request. This ensures that only valid data is processed in your API endpoints, reducing the risk of errors and improving the reliability of your application.

Problem 3: Authentication and Authorization

Implementing authentication and authorization mechanisms is a common challenge in web applications, and FastAPI provides various options to address this problem. However, it can still be complex to implement and manage authentication and authorization in a secure and scalable manner.

One solution is to leverage third-party libraries or frameworks for authentication and authorization, such as OAuth 2.0 or JWT (JSON Web Tokens). These libraries provide standardized protocols and mechanisms for authentication and authorization, reducing the complexity of implementation. Here’s an example using OAuth2 and FastAPI OAuth2 support:


from fastapi import FastAPI
from import OAuth2PasswordBearer

app = FastAPI()

oauth2_scheme = OAuth2PasswordBearer(tokenUrl="/token")

async def get_items(token: str = Depends(oauth2_scheme)):
    # Validate and verify the token
    # Perform authorization checks
    # ...


In the code snippet above, we define an OAuth2 password bearer scheme using the OAuth2PasswordBearer class. This enables authentication and authorization through the token parameter in the get_items endpoint. By specifying the Depends(oauth2_scheme) dependency, FastAPI will handle token validation and verification automatically.

By leveraging these authentication and authorization libraries and integrating them with FastAPI, you can ensure that your application’s endpoints are secure and accessible only to authenticated and authorized users.


A brief introduction to fastapi


FastAPI is a modern, fast, and highly efficient web framework for building APIs with Python. It is designed to provide a high-performance alternative to traditional web frameworks by leveraging asynchronous programming and type annotations. Built on top of Starlette and Pydantic, FastAPI combines the best features of these libraries to deliver a seamless development experience.

One of the key advantages of FastAPI is its performance. It utilizes asynchronous programming techniques, such as async and await, to handle multiple concurrent requests efficiently. This allows FastAPI to handle high loads and deliver excellent performance, making it well-suited for applications that require fast response times and scalability. Additionally, FastAPI utilizes type annotations to automatically generate interactive API documentation. With the help of tools like Swagger UI and ReDoc, developers can easily explore and interact with the API endpoints, making it easier to understand and test the API.

FastAPI also provides strong validation and serialization capabilities through its integration with Pydantic. By defining models using Pydantic’s expressive syntax, developers can specify the expected structure, types, and constraints of the request and response data. FastAPI leverages these models to automatically validate and deserialize incoming requests, reducing the likelihood of data-related errors. Furthermore, FastAPI provides built-in support for various authentication and authorization mechanisms, enabling developers to secure their APIs with ease. Overall, FastAPI’s combination of performance, documentation generation, validation, and security features make it a powerful choice for building high-quality web APIs.


Most popular use cases for fastapi


  1. Building High-Performance APIs: FastAPI excels at building high-performance APIs due to its asynchronous nature and efficient request handling. With the help of async and await keywords, developers can write non-blocking code that can handle multiple concurrent requests. Here’s an example of defining an asynchronous route handler in FastAPI:


from fastapi import FastAPI

app = FastAPI()

async def read_item(item_id: int):
    # Perform asynchronous operations here
    # ...
    return {"item_id": item_id}


  1. Interactive API Documentation: FastAPI provides automatic interactive API documentation generation, making it easier for developers and users to explore and understand the available endpoints, request/response models, and input validation rules. FastAPI leverages the type annotations in Python to generate the documentation, which can be accessed through the Swagger UI or ReDoc. Here’s an example of defining a route with request and response models:


from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI()

class Item(BaseModel):
    name: str
    price: float"/items/")
async def create_item(item: Item):
    # Process the item data here
    # ...
    return {"message": "Item created"}


  1. Integration with Python Ecosystem: FastAPI seamlessly integrates with various Python libraries and frameworks, allowing developers to leverage the existing ecosystem. It works well with ORMs like SQLAlchemy and databases such as PostgreSQL and MongoDB. FastAPI also supports authentication and authorization mechanisms, including OAuth2 and JWT, making it easy to secure API endpoints. Additionally, FastAPI supports dependency injection, allowing for clean and modular code organization. This integration with the Python ecosystem enables developers to build robust and scalable applications.

It’s Really not that Complicated.

You can actually understand what’s going on inside your live applications. It’s a registration form away.

Get Lightrun

Lets Talk!

Looking for more information about Lightrun and debugging?
We’d love to hear from you!
Drop us a line and we’ll get back to you shortly.

By submitting this form, I agree to Lightrun’s Privacy Policy and Terms of Use.