Advanced Features¶

You know the basics: create a throttle, attach it to a route, done. That gets you surprisingly far, but Traffik has a deeper toolkit for the situations where "one size fits all" doesn't quite fit.

This section is for when you're ready to graduate from "100/min" to something that actually mirrors how your API works in the real world.

What's in Here¶

Request Costs¶

Not every request deserves the same weight on the rate limit scale. A health check is a feather. A file upload is a boulder. Request costs let you reflect that asymmetry directly in how quota is consumed.

# A bulk export endpoint that counts as 10 requests worth of quota
throttle = HTTPThrottle("export", rate="100/min", cost=10)

Learn about request costs →

Multiple Rate Limits¶

Production APIs rarely need just one limit. You need a strict burst cap and a generous hourly envelope. Stack throttles to enforce both simultaneously, with a clear failure model.

burst  = HTTPThrottle("api:burst",     rate="20/min")
hourly = HTTPThrottle("api:sustained", rate="500/hour")

# Both must pass — first failure wins
@app.get("/api/data", dependencies=[Depends(burst), Depends(hourly)])
async def get_data():
    ...

Learn about layered throttling →

Exemptions¶

Some clients should never be throttled, such as your internal services, premium users, admin tokens, and whitelisted IPs. The EXEMPTED sentinel lets you carve out those exceptions cleanly, with zero overhead: no counter touched, no backend called.

from traffik import EXEMPTED

async def tiered_identifier(request: Request):
    if request.headers.get("x-admin-token") == ADMIN_TOKEN:
        return EXEMPTED       # Admin? Go right through.
    return request.client.host  # Everyone else: throttled by IP.

Learn about exemptions →

Context-Aware Backends¶

Building a multi-tenant SaaS? Different tenants likely need different backends. Enterprise tenants get their own Redis instance, free-tier users share one pool. The dynamic_backend=True flag makes the throttle resolve which backend to use on every request instead of locking in one backend at startup.

throttle = HTTPThrottle(
    "api:quota",
    rate="1000/hour",
    dynamic_backend=True,   # Resolved per-request from context
)

Learn about context-aware backends →

When Do You Need These?¶