grpc python measure response time

Question

How can I measure the full time grpc-python takes to handle a request? So far the best I can do is:

def Run(self, request, context):
   start = time.time()
   # service code...
   end = time.time()
   return myservice_stub.Response()

But this doesn't measure how much time grpc takes to serialize the request, response, to transfer it over the network.. and so on. I'm looking for a way to "hook" into these steps.

Have you tried the documentation https://grpc.io/docs/languages/python/basics/#request-streaming-rpc — SaaSy Monster, Sep 09 '20 at 11:48
Yes, didn't find. And this link doesn't answer my question, it only measures the python code inside the request handler — user3599803, Sep 09 '20 at 12:00
You can try adding `time.time()` before and after calling `Run(self, request, context` method. — Ajay A, Sep 15 '20 at 08:21
What are you trying to accomplish here? Is this for benchmarking? — gergelykalman, Sep 18 '20 at 16:56
Yes. I'm trying to figure what is the overhead of the grpc serialization, and other things that run before my actual code — user3599803, Sep 21 '20 at 06:35

score 0 · Answer 1 · answered Oct 05 '20 at 19:05

0

You can measure on the client side:

start = time.time()
response = stub.Run(request)
total_end_to_end = time.time() - start

Then you can get the total overhead (serialization, transfer) by reducing the computation of the Run method.

To automate the process, you can add (at least for the sake of the test) the computation time as a field to the myservice_stub.Response.

answered Oct 05 '20 at 19:05

Mark Loyman

1,983
1
14
23

I've already done that, though this will include network + overhead in the server. I hope to measure the overhead in the server alone – user3599803 Oct 06 '20 at 08:39

score 0 · Answer 2 · answered Apr 28 '23 at 08:52

You can use an interceptor for this. the package grpc-interceptors simplifies creating interceptors.

import logging
import time
from typing import Callable, Any
import grpc

class RequestLoggingInterceptor(ServerInterceptor):
    def intercept(
        self,
        method: Callable,
        request: Any,
        context: grpc.ServicerContext,
        method_name: str,
    ) -> Any:
        start_time = time.monotonic()
        ip = context.peer().split(":")[1]

        try:
            start = datetime.now()
            return method(request, context)
        finally:
            request_time = time.monotonic() - start_time
            logging.info(
                f"{ip} -- {datetime.now()} -- {method_name} -- {request_time}s"
            )

grpc python measure response time

2 Answers2