The default SQLite configuration in Django is not ideal for running your application in production. SQLite is optimized for embedded low-concurrency systems out of the box, which is the exact opposite of what your Django application is supposed to do.

Luckily, you can improve concurrency by tweaking a few settings. See how to do it based on the version of Django that you are currently running:

⚠️ WARNING ⚠️

PRAGMA journal_mode=WAL; setting can corrupt your database if used on a network file system (NFS) commonly used by hosting providers like PythonAnywhere. Do not enable it if you are unsure about the type of file storage used.

Make sure to go through SQLite gotchas before using SQLite in production.

In Django 5.1 or newer

In Django 5.1, you can tweak all the necessary changes in your settings.py:

# yourproject/settings.py
DATABASES = {
    "default": {
        "ENGINE": "django.db.backends.sqlite3",
        "OPTIONS": {
            "transaction_mode": "IMMEDIATE",
            "timeout": 5,  # seconds
            "init_command": """
                PRAGMA journal_mode=WAL;
                PRAGMA synchronous=NORMAL;
                PRAGMA mmap_size = 134217728;
                PRAGMA journal_size_limit = 27103364;
                PRAGMA cache_size=2000;
            """,
        },
    }
}

Note that you don’t have to specify PRAGMA foreign_keys = ON because it is set by default by Django itself. You also don’t need to use PRAGMA busy_timeout because you can achieve the same thing by setting the timeout database option.

In Django 5.0, 4.2, or older

1. Enable WAL journal mode

The most impactful change you can make is to enable WAL journal_mode. Without WAL, every write request blocks reads and vice versa, which can kill your throughput.

Enabling WAL mode has no real downsides and can be achieved by running the following command on your database:

sqlite3 db.sqlite3 'PRAGMA journal_mode=WAL;'

You only have to run this command once per database, and the setting will persist.

2. Use IMMEDIATE transactions

Using immediate transactions isn’t a performance improvement. It will decrease your performance when running transactions, but it will avoid unexpected database is locked errors, so it’s worth enabling.

To enable IMMEDIATE transactions, you are going to have to create your database engine:

1. Create a yourproject/sqlite3/base.py file with a DatabaseWrapper class:

   # yourproject/sqlite3/base.py
   from django.db.backends.sqlite3 import base
   
   
   class DatabaseWrapper(base.DatabaseWrapper):
       def _start_transaction_under_autocommit(self):
           # Acquire a write lock immediately for transactions
           self.cursor().execute("BEGIN IMMEDIATE")
   

2. Use the created DatabaseWrapper as your SQLite3 engine in your settings.py:

   
   # yourproject/settings.py
   DATABASES = {
       "default": {
           "ENGINE": "yourproject.sqlite3", # <-- Use our custom engine
           "NAME": BASE_DIR / "db.sqlite3",
       }
   }
   

3. Fine-tune your SQLite settings

A few SQLite settings can improve your application’s performance by a few additional percentage points. The magic values below are now also the default in Rails 7.1 and should give you a good starting point, but feel free to tweak mmap_size, journal_size_limit, and cache_size to best suit your application:

# yourproject/sqlite3/base.py
from sqlite3 import dbapi2 as Database

from django.db.backends.sqlite3 import base
from django.db.backends.sqlite3._functions import register as register_functions
from django.utils.asyncio import async_unsafe


class DatabaseWrapper(base.DatabaseWrapper):
    def _start_transaction_under_autocommit(self):
        # Acquire a write lock immediately for transactions
        self.cursor().execute("BEGIN IMMEDIATE")

    @async_unsafe
    def get_new_connection(self, conn_params):
        conn = Database.connect(**conn_params)
        register_functions(conn)

        conn.execute("PRAGMA foreign_keys = ON")
        # The macOS bundled SQLite defaults legacy_alter_table ON, which
        # prevents atomic table renames.
        conn.execute("PRAGMA legacy_alter_table = OFF")

        conn.execute("PRAGMA journal_mode = WAL")
        conn.execute("PRAGMA synchronous = NORMAL")
        conn.execute("PRAGMA mmap_size = 134217728")
        conn.execute("PRAGMA journal_size_limit = 27103364")
        conn.execute("PRAGMA cache_size = 2000")

        return conn

Fin

That’s it; with these settings, your SQLite database is going to handle the load that most small to medium-sized websites typically get, as long as your use case isn’t write-heavy!