linux grpc測試基於ghz工具

1.測試前提，參考先熟悉grpc測試demo:

參考官網：python 實現grpc client以及service :

https://grpc.io/docs/quickstart/python/

java實現client 和service:https://grpc.io/docs/quickstart/java/

細節我就不講了：

主要說下.proto文件，他是實現grpc的基礎，根據他產生契約helloworld_pb2，以及客戶端服務端調用類：helloworld_pb2_grpc（核心調用client編寫就靠他的Stub 類）：

生成契約和grpc client和 service 類

$ python -m grpc_tools.protoc   -I     ../../protos    --python_out=.      --grpc_python_out=.      ../../protos/helloworld.proto

 

 

 

依賴文件：

grpcio==1.28.1
grpcio-tools==1.28.1
protobuf==3.11.3
coverage>=4.0
cython>=0.29.8
enum34>=1.0.4
#protobuf>=3.5.0.post1
six>=1.10
wheel>=0.29
setuptools==46.1.3

2.linux 環境搭建安裝python3：

虛擬環境：基於virtualenv 

安裝依賴

3.安裝ghz工具，安裝go：

安裝go環境linux教程：https://www.runoob.com/go/go-environment.html

ghza安裝包下載：https://github.com/bojand/ghz/releases 

wget  https://github.com/bojand/ghz/releases/download/v0.52.0/ghz_0.52.0_Linux_x86_64.tar.gz

ghz 幫助文檔：https://ghz.sh/docs/options

配置ghz和go環境變量vi/etc/profile：

export PATH=$PATH:/usr/local/go/bin
export PATH=$PATH:/project

source /etc/profile:

驗證golang ,ghz 環境：

go version 

ghz -h 

4.啓動grpc server:

 

 

 

 編寫一個簡單的服務端grpc_server和對應的client 先mock下server可用，而後進行ghz工具壓測server ：

[root@vm_master helloword]# cat grpc_client.py 
import logging
import time
import grpc
import helloworld_pb2
import helloworld_pb2_grpc
import json

def run():
    with grpc.insecure_channel('localhost:50051') as channel:
        stub = helloworld_pb2_grpc.GreeterStub(channel)
        req={"name":"jack","id":1001}
　　　　 # transfer json  object to bytes to send grpc request 
        body=json.dumps(req).encode("utf-8")
        response = stub.SayHello(helloworld_pb2.HelloRequest(name=body))
        # print(type(response.message))
    print(response.message)



if __name__ == '__main__':
    logging.basicConfig()
    run()

　grpc_server.py:

from concurrent import futures
import logging

import grpc

import helloworld_pb2
import helloworld_pb2_grpc


class Greeter(helloworld_pb2_grpc.GreeterServicer):

    def SayHello(self, request, context):
        return helloworld_pb2.HelloReply(message= request.name)


def serve():
    server = grpc.server(futures.ThreadPoolExecutor(max_workers=10))
    helloworld_pb2_grpc.add_GreeterServicer_to_server(Greeter(), server)
    server.add_insecure_port('[::]:50051')
    server.start()
    server.wait_for_termination()


if __name__ == '__main__':
    logging.basicConfig()
    serve()

　啓動服務端 python grpc_server.py,在運行client測試，查看結果：

C:\Python37\python.exe D:/workspace/demos/python/helloword/grpc_client.py
{"name": "jack", "id": 1001}

Process finished with exit code 0

，接着進行壓測：

前面已經安裝好了ghz ，接着進行壓力測試：

鑑於ghz 命令基於linux因此打算藉助腳本調用 ghz  壓力測試命令，以及請求參數構造：

腳本：

import  subprocess
import json

def exe_cmd(args):


    sub=subprocess.Popen(args,shell=False,stdout=subprocess.PIPE,stderr=subprocess.STDOUT)
    out,err=sub.communicate()

    return out.decode("utf-8")

def demo():
    bytes_data={"name":"zhangsan","id":"10001"}
    bytes_data=json.dumps(bytes_data).encode("utf-8")
    args=["ghz","--skipTLS","--insecure","--proto","../../protos/helloworld.proto",
          "--call","helloworld.Greeter.SayHello","--metadata",bytes_data,"-c","5","--qps","100","-z","1m","--connections","5","-t","20s","0.0.0.0:50051"]
    print(exe_cmd(args))

if __name__ == '__main__':
    demo()

這裏輸出默認是命令行展現結果：

[root@vm_master helloword]# python deso.py

Summary:
Count: 29761
Total: 60.01 s
Slowest: 71.66 ms
Fastest: 0.52 ms
Average: 3.63 ms
Requests/sec: 495.95

Response time histogram:
0.515 [1] |
7.630 [29219] |∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎∎
14.744 [384] |∎
21.859 [133] |
28.973 [8] |
36.088 [2] |
43.202 [0] |
50.317 [3] |
57.431 [2] |
64.546 [4] |
71.660 [1] |

Latency distribution:
10 % in 1.67 ms
25 % in 2.03 ms
50 % in 3.36 ms
75 % in 4.82 ms
90 % in 5.76 ms
95 % in 6.37 ms
99 % in 12.27 ms

Status code distribution:
[OK] 29757 responses
[Unavailable] 2 responses
[Canceled] 2 responses

Error distribution:
[2] rpc error: code = Unavailable desc = transport is closing
[2] rpc error: code = Canceled desc = grpc: the client connection is closing

 爲了方便閱讀，能夠保存爲html 只需添加參數 -o ./result.html -O html既可

args=["ghz","--skipTLS","--insecure","--proto","../../protos/helloworld.proto",
"--call","helloworld.Greeter.SayHello","--metadata",bytes_data,"-o","./result.html","-O","html","-c","5","-n","10","0.0.0.0:50051"]

html參考效果：https://ghz.sh/sample.html

output指標解釋：https://ghz.sh/docs/output

注意事項：

def demo():
bytes_data={"name":"zhangsan","id":"10001"}
bytes_data=json.dumps(bytes_data).encode("utf-8")
args=["ghz","--skipTLS","--insecure","--proto","../../protos/helloworld.proto",
"--call","helloworld.Greeter.SayHello","--metadata",bytes_data,"-c","5","--qps","100","-z","1m","--connections","5","-t","20s","0.0.0.0:50051"]
print(exe_cmd(args))

這裏bytes_data=裏整數按字符串填寫，不然go 轉換出錯，其次 通常 -d {"key":"value"}須要在proto文件定義type,這裏爲了通用傳輸字節，我直接按metadata元數據傳輸他會自動處理

 技術羣答疑羣：