Skip to content
Projects

Tobi-De/django-litestream

View on GitHub

Tobi-De/django-litestream has 26 GitHub stars and 3 forks. GitHot measured 0 stars in 24 hours and 0 in 7 days. Last checked 2026-11-14 06:21 UTC. Last push 2026-11-05.

How these counts are measured

Measured star growth

in 24h
0
in 7 days
0
in 15 days
Collecting history
in 1 month
Collecting history

Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.

About

Django integration with litestream, a standalone streaming replication tool for SQLite.

Data

README.md

Original

django-litestream

PyPI - Version PyPI - Python Version PyPI - Versions from Framework Classifiers MIT License

Django integration for Litestream, the SQLite replication tool. The Litestream binary is bundled in platform-specific wheels — no manual download needed. All upstream commands are available as python manage.py litestream <command>.

Note: This package tracks Litestream's upstream version. The initial release is 0.5.11, matching Litestream v0.5.11. Wrapper-only fixes (no Litestream update) use PEP 440 post-releases like 0.5.11.post1.

Installation

pip install django-litestream

Add to INSTALLED_APPS:

INSTALLED_APPS = [
    "django_litestream",
]

The binary is installed automatically on supported platforms (Linux/macOS x86_64 & arm64). On unsupported platforms, set a custom path:

LITESTREAM = {
    "bin_path": "/usr/local/bin/litestream",
}

Configuration

Configuration is generated dynamically from Django settings — no YAML file needed.

# settings.py
LITESTREAM = {
    "path_prefix": "",      # prepended to S3 replica paths
    "bin_path": "...",      # custom binary path (optional)
    "dbs": [                # databases to replicate
        {"path": "default"},
    ],
}

All Litestream config options are supported (addr, logging, mcp-addr, access-key-id, shutdown-timeout, etc.) and forwarded verbatim; only path_prefix, bin_path and vfs are consumed by the wrapper.

Database aliases

Use Django database aliases instead of file paths wherever db_path is required:

python manage.py litestream restore default        # instead of /path/to/db.sqlite3
python manage.py litestream sync default

Auto-generated replicas

If no replica is specified for a database, an S3 replica is auto-generated using environment variables:

LITESTREAM = {
    "dbs": [{"path": "default"}],
}
# Uses: LITESTREAM_REPLICA_BUCKET, LITESTREAM_ACCESS_KEY_ID, LITESTREAM_SECRET_ACCESS_KEY

Explicit replica config:

LITESTREAM = {
    "dbs": [
        {
            "path": "default",
            "replica": {
                "type": "s3",
                "bucket": "my-bucket",
                "path": "db.sqlite3",
            },
        },
    ],
}

Multiple databases with path_prefix

DATABASES = {
    "default": {"ENGINE": "django.db.backends.sqlite3", "NAME": BASE_DIR / "db.sqlite3"},
    "cache":  {"ENGINE": "django.db.backends.sqlite3", "NAME": BASE_DIR / "cache.sqlite3"},
}

LITESTREAM = {
    "path_prefix": "myproject",
    "dbs": [
        {"path": "default"},           # → myproject/db.sqlite3
        {
            "path": "cache",
            "replica": {
                "type": "s3",
                "bucket": "my-cache-bucket",
                "path": "cache.sqlite3",
            },
        },
    ],
}

Per-database options

Every database-level option is passed through untouched; the wrapper only resolves the path alias and fills in an S3 replica when neither replica nor replicas is set:

LITESTREAM = {
    "dbs": [
        {
            "path": "default",
            "snapshot": {"interval": "6h", "retention": "168h"},
            "monitor-interval": "5s",
            "checkpoint-interval": "1m",
        },
    ],
}

dir-based entries (directory watching) are forwarded as-is, so pattern, recursive, watch, meta-dir, etc. work too. Durations use Go syntax (6h, 168h, 30m), not 7d/1w.

Custom commands

These commands are unique to django-litestream — they don't exist in upstream Litestream.

litestream config

Display the generated configuration in YAML format.

python manage.py litestream config

Useful for verifying your setup before running replication.

litestream verify

Checks backup integrity end-to-end: writes a verification row, waits for replication, restores the latest backup, and confirms the row exists.

python manage.py litestream verify default

Upstream commands

All Litestream commands are exposed as management commands. Run python manage.py litestream to see the full list. Key notes:

  • replicate: Run in production via a process manager. Use -exec to wrap your application server.
  • restore: Supports -timestamp, -generation, -index for point-in-time recovery.
  • Daemon control (info, list, register, unregister, start, stop): Communicate with a running replicate daemon over IPC.

The only difference from upstream is that Django database aliases work anywhere db_path is expected.

VFS Read Replicas

Read-only access to cloud-stored replicas without downloading the entire database. Pages are fetched on-demand.

pip install django-litestream[vfs]

This installs django-litestream, the django-litestream-vfs Django integration, and the upstream litestream-vfs package, which bundles the compiled VFS extension.

Setup

# settings.py
from django_litestream_vfs import get_vfs_databases

LITESTREAM = {
    "dbs": [{"path": "default"}],
    "vfs": {
        "prod_replica": "s3://mybucket/db.sqlite3",
    },
}

DATABASES = {
    "default": {
        "ENGINE": "django.db.backends.sqlite3",
        "NAME": BASE_DIR / "db.sqlite3",
    },
    **get_vfs_databases(),  # adds VFS replicas
}

Usage

User.objects.using("prod_replica").all()  # read-only, on-demand page fetch

VFS replicas are read-only. Only x86_64 and arm64 supported. Multiple storage backends supported (S3, GCS, Azure Blob).

Note: the upstream litestream-vfs package currently ships wheels for Python 3.8–3.13. On Python 3.14 the VFS extra will not install until upstream publishes a cp314 wheel.

License

MIT. The bundled Litestream binary uses the Apache 2.0 license.

Explore related projects