Skip to main content

ASGI compliant web server

Project description


An ASGI compliant web server the goal of which is to support multiple asynchronous models. This started as a companion project to the Tonberry ASGI framework.

1. Post-Fork Multiprocessing

The first strategy is a post-fork model creating a new process for every request. This is the only model available right now.

2. Async/Await

The second strategy is one that a lot of current ASGI servers already do. In a single process just handle requests using asyncio/uvloop.

3. Pre-Fork Multiprocessing with Async/Await

Start up a pool of processes which can share a socket and each take a set number of requests and then handle them internally with async.


Just do the usual

$ pip install qactuar


During the install it creates a command line app.

Command Line

$ qactuar module:app

If your module is in the python path then it will get imported and any ASGI compatible object in the module can be called. If the apps are setup in the config (see Configuration section below) then you can just start up qactuar without any arguments.

$ qactuar


from tonberry import create_app, expose
from tonberry.content_types import TextPlain

import qactuar

class Root:
    async def index(self) -> TextPlain:
        return "Hello, world!"

if __name__ == "__main__":
    app = create_app(Root)
    # other keyword arguments for run() are host, port and conf

For a nice ASGI framework, check out Tonberry!


File and programatic based configurations are supported. For a config file just create an environment variable QACTUAR_CONFIG and set the value to the absolute or relative path of a JSON file. This file can overwride any of the default values listed here. Only the values you wish to override need to be provided.

  • HOST: str = "localhost"
  • PORT: int = 8000
  • ADMIN_HOST: str = "localhost"
  • ADMIN_PORT: int = 1986
  • SELECT_SLEEP_TIME: float = 0.025
  • RECV_TIMEOUT: float = 0.001
  • RECV_BYTES: int = 65536
  • MAX_PROCESSES: int = 500
  • GATHER_PROC_STATS: bool = False
  • REQUEST_TIMEOUT: float = 5
  • APPS: Dict[str, str] = {}
  • LOGS: Dict[str, Any] = default_log_setup (see below)

The APPS dictionary takes a route as the key and a module:app style string as the value. Multiple applications can be hosted at the same time by registering each at its own route. A basic example can be seen in the qactuar_config.json file.

The Config dataclass

The config is managed in a dataclass object and can be created programmatically. All arguments are optional and are defined above.

config = qactuar.Config(HOST=''), conf=config)

The LOGS dictionary is for the logging.config setup as detailed in the Python documentation here. It uses logging.config.dictConfig to set the logging configs. The loggers used throught the code are qt_server (used by the parent process), qt_child (used in the child processes), qt_access (used to log each request), qt_exception (used to log any exceptions)

Default Log Setup

    "version": 1,
    "disable_existing_loggers": false,
    "formatters": {
        "standard": {
            "format": "{asctime} {levelname} {message}",
            "style": "{"
        "access": {
            "format": "{asctime} ACCESS {message}",
            "style": "{"
    "handlers": {
        "console": {
            "class": "logging.StreamHandler",
            "level": "DEBUG",
            "formatter": "standard",
            "stream": "ext://sys.stdout"
        "access": {
            "class": "logging.StreamHandler",
            "level": "INFO",
            "formatter": "access",
            "stream": "ext://sys.stdout"
        "exception": {
            "class": "logging.StreamHandler",
            "level": "ERROR",
            "formatter": "standard",
            "stream": "ext://sys.stderr"
    "loggers": {
        "qt_server": {
            "handlers": ["console"],
            "level": "DEBUG"
        "qt_child": {
            "handlers": ["console"],
            "level": "DEBUG"
        "qt_access": {
            "handlers": ["access"],
            "level": "INFO"
        "qt_exception": {
            "handlers": ["exception"],
            "level": "ERROR"

If changing the LOGS config then the whole dictionary needs to be replaced. Individual parts of the log config cannot be changed by themselves.

Future Config Options

  • SSL_CERT_PATH: str = ""
  • SSL_KEY_PATH: str = ""

Tornado Apps

Included is a utility wrapper to take a Tornado Request Handler and make it work with ASGI. See for an example.


The plan is to support an extra socket connection that can accept connections for adminitrative purposes. Maybe for viewing system recources, viewing the processes and current load, changing configs on the fly, installing new apps via git clone without restarting the server. This is still in the works.


Please read for details on our code of conduct, and the process for submitting pull requests.


SemVer is used for versioning. For the versions available, see the tags on this repository.



This project is licensed under the MIT License - see the LICENSE.txt file for details


  • Admin socket
  • UPD support
  • WebSockets
  • Send http.disconnect to app when each socket closes
  • Filter HTTP/2-3 pseudo headers
  • Client streaming, check "more_body" in send method
  • Async only model
  • Pre-Fork model
  • TESTS!!!
  • Docs

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for Qactuar, version 0.1.0
Filename, size File type Python version Upload date Hashes
Filename, size Qactuar-0.1.0.tar.gz (16.4 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page