Added flexible customization of gunicorn via environment variables #2813
Conversation
There was a problem hiding this comment.
If we implement the functionality of changing the configuration parameters of gunicorn, then it seems to me that it is necessary to correct/add the following:
- The documentation of the
wemake-django-templateproject - In my opinion, the parameters in
.envshould be stored not as a dictionary, but as a set, so it will be consistent. _load_gunicorn_settings- must receive keys from.envwith the prefixGUNICORN_
@sobolevn Do you think it's worth discussing?
| 'loglevel': 'info', | ||
| 'preload_app': False, | ||
| 'wsgi_app': 'server.wsgi.application', | ||
| 'workers': 1, |
There was a problem hiding this comment.
I think it's better to use multiprocessing.cpu_count() * 2 + 1
There was a problem hiding this comment.
A valid point. Thanks. Waiting for @sobolevn feedback.
| ) | ||
|
|
||
|
|
||
| GUNICORN_WSGI_DEFAULTS: MappingProxyType[str, Any] = MappingProxyType({ |
There was a problem hiding this comment.
Maybe it's better to use a structure with attributes and load default values using getattr()
There was a problem hiding this comment.
Sorry, the use-case is not really clear for me. Why do you want to have different gunicorn settings? They always must be the same. Otherwise, different envs might behave in a different way.
Also question: how do you want env change to work? I will still have to reload gunicorn for changes to make effect.
| 'loglevel': 'info', | ||
| 'preload_app': False, | ||
| 'wsgi_app': 'server.wsgi.application', | ||
| 'workers': 1, |
Co-authored-by: sobolevn <mail@sobolevn.me>
@sobolevn The idea was more along the following lines: to make it easier to adjust Gunicorn settings in production without creating a pull request and editing config files. Just restart the deployment job in CI/Kubernetes pod (or let Kube itself update the env after some time). Perhaps such tricks are excessive for this template. Still, I wanted to suggest it—maybe something useful could come out of it. :) |
I am not a contributor, but I agree with the point. |
| ) | ||
|
|
||
| # https://docs.gunicorn.org/en/22.0.0/settings.html#bind | ||
| bind = GUNICORN_WSGI_SETTINGS.get('bind', GUNICORN_WSGI_DEFAULTS['bind']) |
There was a problem hiding this comment.
Question / idea. Maybe we can use module level __getattr__ to get needed attributes?
Added flexible customization of gunicorn via environment variables.
1. Flexibility and Environment-Based Configuration
Allows dynamic parameter changes via the GUNICORN_WSGI_SETTINGS environment variable, making it suitable for different environments (dev/stage/prod) and Docker deployments.
2. Comprehensive Configuration
Includes all major Gunicorn parameters with clear default values (GUNICORN_WSGI_DEFAULTS).
3. Documentation and Clarity
Detailed comments with links to Gunicorn’s official documentation.
Usage examples and parameter explanations.
4. Advanced Worker Settings
Fine-tuned worker management (worker_class, threads, worker_connections).
Logging customization (access_log_format, loglevel).
Timeout controls (timeout, graceful_timeout, keepalive).
The second file lacks these options.
5. Production Optimization
Supports preload_app, max_requests_jitter, and worker_tmp_dir (for tmpfs).