af = AsyncFetch({
    'first': AsyncFetch.mk_task(build_url('request-info')),
    'second': AsyncFetch.mk_task('http://example.com/'),
    'fail': AsyncFetch.mk_task(build_url('404'))
})
responses = af.go()

:param fail_silently: bool, do not raise exceptions, default is False;
    for test purpose, do not use in production
:param num_retries: int, *optional*, default is -1; -1 - no retries; 0 - use AsyncFetch.num_retries
:param autodetect_content_type: if no `content-type` header was specified, set `content-type` as
    `application/json` for dict, and `text/html` otherwise; default is True
:param json_encoder: JSONEncoder, *optional*, JSON encoder for data serialization
    tries to use DRF's encoder, or default JSONEncoder from json package; default is JSONEncoder
:param url: str, *required*, url address
:param api_key: str, optional API key passed into HEADERS dict
:param data: dict, *optional*, request data. Default is None,
:param method: str, *optional*, HTTP request method. Default is True.
:param headers: dict, *optional*, optional HTTP headers
:param response_type: str, *optional*, HTTP response type
    (in fact it's just aiohttp's method name, i.e. text, or json); default is 'json'
:param language_code: str, set `accept-language` header
:param timeout: float, *optional*, time to wait for response in seconds before TimeoutError
:param query: dict, *optional*, url get arguments
:param do_not_wait: bool, *optional*, fail silently with no retries and empty resultset

af0 = AsyncFetch({})
tcp_connector = af0.get_tcp_connector()

af1 = AsyncFetch({
    '1': AsyncFetch.mk_task(build_url('request-info')),
    '2': AsyncFetch.mk_task(build_url('request-info')),
}, tcp_connector=tcp_connector)
responses = af1.go()