kserve/docs/samples/v1beta1/transformer/auth-forwarding.md

Credential Forwarding for Transformers

The transformer must forward the caller's credentials (the Authorization header) to the predictor so that the predictor's auth proxy (e.g. kube-rbac-proxy) can validate the original caller's identity.

SDK-Based Transformers (Automatic)

If your transformer extends kserve.Model and uses the built-in predict() method (i.e., you override preprocess and/or postprocess but not predict), the KServe SDK automatically forwards the Authorization header from the incoming request to the predictor.

No additional configuration is needed. The following headers are forwarded automatically:

Header	Purpose
Authorization	Caller credentials for predictor auth
x-request-id	Request tracing
x-b3-traceid	Distributed tracing (B3/Zipkin)

Example

from kserve import Model, InferRequest, InferResponse

class MyTransformer(Model):
    def __init__(self, name: str):
        super().__init__(name)
        self.ready = True

    def preprocess(self, payload: InferRequest, headers=None) -> InferRequest:
        # Transform input - Authorization header is forwarded automatically
        # when the SDK calls predict() on the predictor
        return payload

    def postprocess(self, result: InferResponse, headers=None) -> InferResponse:
        return result

Custom Transformers (Manual Forwarding Required)

If your transformer overrides the predict() method or makes its own HTTP calls to the predictor, you are responsible for forwarding the Authorization header yourself.

What You Need To Do

1. Accept the headers parameter in your predict method
2. Extract the Authorization header from the incoming headers
3. Include it in the request you send to the predictor

Example

import httpx
from kserve import Model

class MyCustomTransformer(Model):
    def __init__(self, name: str):
        super().__init__(name)
        self.ready = True

    async def predict(self, payload, headers=None, response_headers=None):
        # Build headers for the predictor call
        predict_headers = {"Content-Type": "application/json"}

        # Forward the Authorization header for auth-enabled predictors
        if headers and "authorization" in headers:
            predict_headers["authorization"] = headers["authorization"]

        # Make the call to the predictor
        async with httpx.AsyncClient() as client:
            response = await client.post(
                f"http://{self.predictor_host}/v1/models/{self.name}:predict",
                json=payload,
                headers=predict_headers,
            )
            return response.json()

Headers You Should Forward

At minimum, forward the Authorization header. For full compatibility with KServe observability, also forward:

• x-request-id - for request tracing
• x-b3-traceid - for distributed tracing

Important Notes

• The predictor authorizes the end user, not the transformer service account. The transformer is a pass-through for credentials.
• The Host header from the incoming request should not be forwarded - it contains the transformer's hostname, not the predictor's.
• When auth is enabled, the predictor's kube-rbac-proxy validates the bearer token in the Authorization header against Kubernetes RBAC.