last = defaultLogBufferCount
	}

	lastN = make([]log.Info, last)
	sys.RLock()
	sys.logBuf.Do(func(p any) {
		if p != nil {
			lg, ok := p.(log.Info)
			if ok && lg.SendLog(node, logKind) {
				lastN[cnt%last] = lg
				cnt++
			}
		}
	})
	sys.RUnlock()
	// send last n console log messages in order filtered by node
	if cnt > 0 {
		for i := range last {

https://github.com/google/guava/wiki/HowToContribute#code-contributions
and
https://github.com/google/guava/blob/master/CONTRIBUTING.md
before sending a pull request.

We generally welcome PRs for fixing trivial bugs or typos, but please refrain
from sending a PR with significant changes unless explicitly requested.

  @Override public Response intercept(Interceptor.Chain chain) throws IOException {
    Request request = chain.request();

    long t1 = System.nanoTime();
    logger.info(String.format("Sending request %s on %s%n%s",
        request.url(), chain.connection(), request.headers()));

    Response response = chain.proceed(request);

    long t2 = System.nanoTime();

This includes, for example:

* Email notifications sent after performing an action:
    * As connecting to an email server and sending an email tends to be "slow" (several seconds), you can return the response right away and send the email notification in the background.
* Processing data:

## Base64 vs Files { #base64-vs-files }

Consider first if you can use [Request Files](../tutorial/request-files.md) for uploading binary data and [Custom Response - FileResponse](./custom-response.md#fileresponse--fileresponse-) for sending binary data, instead of encoding it in JSON.

JSON can only contain UTF-8 encoded strings, so it can't contain raw bytes.

{* ../../docs_src/dependencies/tutorial013_an_py310.py ln[19:21] *}

...would be run after the response finishes sending the slow data:

{* ../../docs_src/dependencies/tutorial013_an_py310.py ln[30:38] hl[31:33] *}

But as `generate_stream()` doesn't use the database session, it is not really necessary to keep the session open while sending the response.

It enables Brotli support in addition to tranparent Gzip support,
provided Accept-Encoding is not set previously.  Modern web servers
must choose to return Brotli responses.  n.b. It is not used for
sending requests.

```java
OkHttpClient client = new OkHttpClient.Builder()
  .addInterceptor(BrotliInterceptor.INSTANCE)
  .build();
```

```kotlin
implementation("com.squareup.okhttp3:okhttp-brotli:5.3.0")
```

Being able to use asynchronous functions in your tests could be useful, for example, when you're querying your database asynchronously. Imagine you want to test sending requests to your FastAPI application and then verify that your backend successfully wrote the correct data in the database, while using an async database library.

Let's look at how we can make that work.

# OpenAPI Webhooks { #openapi-webhooks }

There are cases where you want to tell your API **users** that your app could call *their* app (sending a request) with some data, normally to **notify** of some type of **event**.

This means that instead of the normal process of your users sending requests to your API, it's **your API** (or your app) that could **send requests to their system** (to their API, their app).

This is normally called a **webhook**.

/// info

Added in FastAPI 0.134.0.

///

## What is a Stream? { #what-is-a-stream }

"**Streaming**" data means that your app will start sending data items to the client without waiting for the entire sequence of items to be ready.

So, it will send the first item, the client will receive and start processing it, and you might still be producing the next item.

```mermaid