Deployment

---

Choosing Your Server

Use Case	Recommended Server	Why
Sync views only (def)	Gunicorn (WSGI)	Mature, stable, simple
Async views (async def)	Uvicorn, Daphne (ASGI)	Native async support, better concurrency
Mixed sync/async	Uvicorn (ASGI)	Handles both seamlessly

Not sure?

Start with Uvicorn — it supports both sync and async views, and is the recommended default for django-flosse v0.2.0+.

---

ASGI — Uvicorn (Recommended for Async)

For best performance with async def SSE views, run your project with an ASGI server:

uvicorn myproject.asgi:application \
    --workers 4 \
    --loop uvloop \
    --http h11

Why these flags?

• --workers: Scale horizontally (one per CPU core)
• --loop uvloop: Faster event loop (optional but recommended)
• --http h11: Better SSE compatibility with some proxies

Production Example (with Gunicorn + Uvicorn workers)

gunicorn myproject.asgi:application \
    -k uvicorn.workers.UvicornWorker \
    --workers 4 \
    --worker-connections 1000 \
    --timeout 120

This combines Gunicorn's process management with Uvicorn's async performance.

---

WSGI — Gunicorn (Sync Only)

Each SSE connection holds one WSGI worker for its entire duration. Use gthread workers so each worker can serve multiple streams concurrently:

gunicorn myproject.wsgi:application \
    --worker-class gthread \
    --workers 2 \
    --threads 4

This allows up to workers × threads = 8 concurrent SSE streams per server.

Tuning

Start with --workers $(nproc) and --threads 4. Increase --threads if most connections are idle (SSE is I/O-bound).

Async views on WSGI

Async generators will run in a thread pool on WSGI servers, negating most performance benefits. Use ASGI for async def views.

---

WSGI — uWSGI

[uwsgi]
module        = myproject.wsgi:application
workers       = 2
threads       = 4
enable-threads = true

---

Nginx

X-Accel-Buffering: no is set automatically by @sse_stream — no extra Nginx config is required for buffering.

For long-lived connections, increase the read timeout:

location /sse/ {
    proxy_pass             http://django;
    proxy_read_timeout     3600;
    proxy_buffering        off;
    proxy_cache            off;
}

Cloudflare / CDN users

Add these headers to prevent proxy buffering:

proxy_set_header Connection '';
proxy_http_version 1.1;
chunked_transfer_encoding off;

---

Compatibility

django-flosse	Django	Python	Server Support
0.1.x	3.2 – 6.0	3.9 – 3.14	WSGI only
0.2.0+	4.2 – 6.0	3.9 – 3.14	WSGI + ASGI

Async support is here!

Native async def generator views are available in v0.2.0+. Use an ASGI server (Uvicorn, Daphne) for best performance. See Async Support Guide for details.

---

Monitoring & Health Checks

Log Stream Lifecycle

# settings.py
LOGGING = {
    "version": 1,
    "handlers": {
        "console": {"class": "logging.StreamHandler"},
    },
    "loggers": {
        "django_flosse": {
            "handlers": ["console"],
            "level": "INFO",
        },
    },
}

Simple Health Endpoint

# views.py
from django.http import JsonResponse

def sse_health(request):
    return JsonResponse({
        "status": "ok",
        "sse_support": True,
        "async_ready": True,  # v0.2.0+
    })

---

Troubleshooting

Issue	Solution
Connections drop after ~60s	Disable proxy buffering; increase proxy_read_timeout
Async views run slowly	Ensure you're using ASGI (Uvicorn), not WSGI
High memory usage per worker	Reduce --threads (WSGI) or --workers (ASGI); monitor with psutil
CancelledError in logs	Normal on disconnect; handle gracefully or suppress
SSE not working behind Cloudflare	Add proxy_set_header Connection '' + disable caching