communities:
                description: The BGP communities to be associated with the announcement.
                  Each item can be a standard community of the form 1234:1234, a large
                  community of the form large:1234:1234:1234 or the name of an alias
                  defined in the Community CRD.
                items:
                  type: string
                type: array

* It uses a "spooled" file:
    * A file stored in memory up to a maximum size limit, and after passing this limit it will be stored in disk.
* This means that it will work well for large files like images, videos, large binaries, etc. without consuming all the memory.
* You can get metadata from the uploaded file.

        response = client.post("/files/", files={"file": file})
    assert response.status_code == 200, response.text
    assert response.json() == {"file_size": 14}


@needs_py39
def test_post_large_file(tmp_path, client: TestClient):
    default_pydantic_max_size = 2**16
    path = tmp_path / "test.txt"
    path.write_bytes(b"x" * (default_pydantic_max_size + 1))

    with path.open("rb") as file:

// A chunk has the following format:
//
//	<chunk-size-as-hex> + ";chunk-signature=" + <signature-as-hex> + "\r\n" + <payload> + "\r\n"
//
// First, we read the chunk size but fail if it is larger
// than 16 MiB. We must not accept arbitrary large chunks.
// One 16 MiB is a reasonable max limit.
//
// Then we read the signature and payload data. We compute the SHA256 checksum

	n := len(b) * count

	if len(b) == 0 {
		return []byte{}
	}

	// Past a certain chunk size it is counterproductive to use
	// larger chunks as the source of the write, as when the source
	// is too large we are basically just thrashing the CPU D-cache.
	// So if the result length is larger than an empirically-found
	// limit (8KB), we stop growing the source string once the limit

And multiple processes normally **don't share any memory**. This means that each running process has its own things, variables, and memory. And if you are consuming a large amount of memory in your code, **each process** will consume an equivalent amount of memory.

   *
   * <p><b>Note:</b> coalescing requires calling {@code .equals()} on any connected values, which
   * may be expensive depending on the value type. Using this method on range maps with large values
   * such as {@link Collection} types is discouraged.
   *
   * @since 22.0
   */
  void putCoalescing(Range<K> range, V value);

Since the domain certificates are stored in the gateway namespace, it is recommended to keep each
gateway in a dedicated namespace and restrict access.

For large-scale gateways it is optionally possible to use a dedicated pilot in the gateway namespace.

### Additional test templates

A number of helm test setups are general-purpose and should be installable in any cluster, to confirm

 *     process(part.headers, part.body)
 *   }
 * }
 * ```
 *
 * Note that [nextPart] will skip any unprocessed data from the preceding part. If the preceding
 * part is particularly large or if the underlying source is particularly slow, the [nextPart] call
 * may be slow!
 *
 * Closing a part **does not** close this multipart reader; callers must explicitly close this with
 * [close].
 *

    with path.open("rb") as file:
        response = client.post("/files/", files={"file": file})
    assert response.status_code == 200, response.text
    assert response.json() == {"file_size": 14}


def test_post_large_file(tmp_path):
    default_pydantic_max_size = 2**16
    path = tmp_path / "test.txt"
    path.write_bytes(b"x" * (default_pydantic_max_size + 1))

    client = TestClient(app)
    with path.open("rb") as file: