This page guides you through the steps of configuring your Starlette application to work with Apitally. If you don’t have an account yet, now would be a good time to sign up.

Once you’re done with this guide, you will be able to:

  • Analyze traffic to your application
  • Keep track of errors, response times and payload sizes
  • Understand how individual API consumers use your application
  • Log and inspect each API request and response
  • Monitor your application’s uptime and receive alerts when it’s down
  • Set up custom alert rules and get notified when they’re triggered

​
Create app

To get started, create a new app in the Apitally dashboard and select as your framework.

Here you can also configure the environments (e.g. prod and dev) for your app, or simply accept the defaults.

Take a note of the client ID displayed after submitting. You will need it in the next step.

​
Add middleware

Next, install the Apitally client library in your project with the extra.

Add the Apitally middleware to your Starlette application and provide the client_id for your app. You’ll find the client_id on the Setup instructions page for your app in the Apitally dashboard, which is displayed immediately after creating the app.

from starlette.applications import Starlette
from apitally.starlette import ApitallyMiddleware

app = Starlette(routes=[...])
app.add_middleware(
    ApitallyMiddleware,
    client_id="your-client-id",
    env="dev",  # or "prod" etc.
)

If you're also using other middlewares, add the ApitallyMiddleware last, so that it wraps the existing stack of middlewares.

Deploy your application with these changes, or restart if you're testing locally.

At this point the basic setup for your application is complete and you will start seeing data in the Apitally dashboard. However, you won't be able to analyze API consumers yet.

​
Identify consumers

In order to analyze API traffic by consumers in Apitally, your application must identify which consumers requests are coming from.

How you identify API consumers depends on your application and use case. If your application uses authentication, it would make sense to use the authenticated identity (e.g. username) as the consumer identifier.

To associate requests with consumers, provide a to the ApitallyMiddleware that takes a Request object as an argument and returns a consumer identifier or None.

Alternatively, you can set the apitally_consumer attribute on the request.state object, for example in another middleware or directly in your endpoint functions.

The consumer identifier should be a string (max. 128 characters long) that uniquely identifies the consumer of the API, e.g. customer-123. Once the first request is received from a consumer, it will show up in the Apitally dashboard.

Now the Consumers page in the Apitally dashboard shows you information about all consumers that have made requests to your application. You can also filter insights on other the dashboards by consumer and better understand how each of them use your application.

​
Name and group consumers

By default, Apitally generates a name for new consumers from the consumer identifier. You can always change it in the Apitally dashboard, and you can assign consumers to groups there too.

If you have many consumers, you may prefer to set a name and group programmatically from your application.

from starlette.requests import Request
from apitally.starlette import ApitallyConsumer

def identify_consumer(request: Request) -> ApitallyConsumer | None:
    if request.user.is_authenticated:
        return ApitallyConsumer(
            identifier=request.user.identity,
            name=request.user.display_name,
            group="Customers",
        )
    return None

​
Configure request logging

The request log is an opt-in feature available on our paid plans only.

Logging of individual requests and responses is disabled by default to protect potentially sensitive data. You can enable it by providing a RequestLoggingConfig object, which allows you to configure in detail what parts of the request and response should be logged. There are also options for masking sensitive information and excluding certain requests from logging.

NestJS