Customizing uWSGI Configuration

MrDoc provides a default uWSGI configuration optimized for mid-range server setups. This configuration is generally suitable for servers with:

* 4 to 8 GB of RAM (each worker process limited to 2 GB, after which it will automatically restart);
* 4 CPU cores (worker process count limited to 4);
* Moderate traffic load (maximum requests per worker set to 1000);

If the default uWSGI config does not work well on your server, for example:

* On entry-level devices, the CPU or memory usage is too high;
* On mid-to-high performance servers with heavy traffic, the site response is unsatisfactory or system resources are underutilized;

You can customize the uWSGI configuration file according to your hardware and performance needs.

---

### Steps to customize uWSGI config:

1. Inside the MrDoc project directory, create a new file named `uwsgi_custom.ini` under the `/config/` folder.

2. Copy the contents of the existing `/config/uwsgi.ini` file into your new `uwsgi_custom.ini`.

Example default configuration:

```ini
[uwsgi]
master = true                           # Enable master process mode to manage worker processes
processes = 4                           # Spawn 4 worker processes (set according to CPU cores)
chdir = /app/MrDoc                      # Set project root directory path
wsgi-file= /app/MrDoc/MrDocPro/wsgi.py  # Set WSGI application entry point
pidfile = /app/MrDoc/config/uwsgi.pid   # Path to store the process ID file
http = 0.0.0.0:10086                    # Enable HTTP server listening on all IPs at port 10086
logto = /app/MrDoc/log/uwsgi_log.log    # Log file output path
log-maxsize = 3000000                   # Max log file size 3MB before rotation
chmod-socket = 660                      # Socket file permissions
vacuum = true                           # Clean up environment (remove socket and pid files) on shutdown
master = true                           # Duplicate setting, can be removed
enable-threads = true                   # Enable thread support
max-requests = 1000                     # Restart worker after handling 1000 requests (prevents memory leaks)
reload-on-as = 2048                     # Restart worker if virtual memory usage exceeds 2048MB
reload-on-rss = 2048                    # Restart worker if physical memory usage exceeds 2048MB
evil-reload-on-as = 2048                # Force restart ignoring active requests if virtual memory > 2048MB
evil-reload-on-rss = 2048               # Force restart ignoring active requests if physical memory > 2048MB
buffer-size = 65536                     # Internal buffer size set to 64KB (for large requests)
http-timeout = 300                      # HTTP request timeout in seconds
socket-timeout = 300                    # Socket connection timeout in seconds
reload-mercy = 1                        # Grace period (seconds) for worker to finish active requests on restart
worker-reload-mercy = 1                 # Grace period for worker restart to complete ongoing requests
```

---

Each option includes comments explaining its purpose. You can adjust the parameters according to your server’s actual specifications and workload.