grpc同时提供restful api

概述

根据google cloud的API设计指南：
对比rest api和rpc： 2010 年，大约 74% 的公共网络 API 是 HTTP REST，虽然 HTTP/JSON API 在互联网上非常流行，但它们承载的流量比传统的 RPC API 要小，视频内容，数据中心内部 RPC API来承载大多数网络流量
在实际使用中，人们会出于不同目的选择 RPC API 和 HTTP/JSON API，理想情况下，API 平台应该为所有类型的 API 提供最佳支持

而藏宝阁的使用场景，服务端调用的接口如果统一使用rpc， 也可能存在需要提供rest api的可能，比如同时提供给游戏/大神/cc等调用，grpc不一定方便对接

Google API 和 Cloud Endpoints gRPC API 使用 HTTP 映射功能进行 JSON/HTTP 到 Protocol Buffers/RPC 的转码： gRPC Transcoding

gRPC Transcoding的映射定义demo, 留意option(google.api.http)配置:

// Returns a specific bookstore shelf.
rpc GetShelf(GetShelfRequest) returns (Shelf) {
  // Client example - returns the first shelf:
  //   curl http://DOMAIN_NAME/v1/shelves/1
  option (google.api.http) = { get: "/v1/shelves/{shelf}" };
}

...
// Request message for GetShelf method.
message GetShelfRequest {
  // The ID of the shelf resource to retrieve.
  int64 shelf = 1;
}

也可以使用独立的yaml文件配置映射，不过推荐在proto文件中定义

支持gRPC Transcoding的系统包括：Google APIs, Cloud Endpoints, gRPC Gateway, and Envoy proxy

Google APIs我们可以无视, 下面对其他3种方式做个说明

Cloud Endpoints

Cloud Endpoints是google cloud提供的服务，部署结构如下图所示，

image.png

留意图中红框的部分是部署在google cloud的Endpoints服务，因为依赖google cloud提供的服务，我们也无法使用

gRPC Gateway

使用这个grpc服务来举例: https://github.com/GoogleCloudPlatform/python-docs-samples/blob/main/endpoints/bookstore-grpc

如下命令启动gprc server， 默认端口8000

python bookstore_server.py

gRPC Gateway 完全是go的实现，以go grpc为基础，需要生成go grpc的stub，在额外生成grpc gateway的stub
所以需要安装go的环境, 安装protobuf
之后 go install:
google.golang.org/grpc/cmd/protoc-gen-go-grpc@latest
google.golang.org/protobuf/cmd/protoc-gen-go@latest

在 bookstore-grpc 创建gw目录，生成go gpc/grpc gateway stub及编写go代码

mkdir gw
cd gw
go mod init cbg/demorpc/gw

使用 http_bookstore.proto, 生成stub代码
留意需要需要修改下http_bookstore.proto, 加上option go_package = "cbg/demorpc/gw/pb";
执行如下命令

cp ../http_bookstore.proto ./
# 留意修改加上option go_package = "cbg/demorpc/gw/pb";

# 因为http_bookstore.proto依赖googleapi的proto file
git clone https://github.com/googleapis/googleapis.git

# 生成stub代码
# mkdir pb
protoc -I . -I ./googleapis/ --grpc-gateway_out ./pb \
    --grpc-gateway_opt paths=source_relative \
    --grpc-gateway_opt generate_unbound_methods=true \
    --go_out ./pb --go_opt paths=source_relative \
    --go-grpc_out ./pb --go-grpc_opt paths=source_relative \
    http_bookstore.proto

创建main.go文件，内容如下:

package main

import (
  "context"
  "flag"
  "net/http"

  "github.com/golang/glog"
  "github.com/grpc-ecosystem/grpc-gateway/v2/runtime"
  "google.golang.org/grpc"
  "google.golang.org/grpc/credentials/insecure"

  pb "cbg/demorpc/gw/pb"
)

var (
  // command-line options:
  // gRPC server endpoint
  grpcServerEndpoint = flag.String("grpc-server-endpoint",  "localhost:8000", "gRPC server endpoint")
)

func run() error {
  ctx := context.Background()
  ctx, cancel := context.WithCancel(ctx)
  defer cancel()

  // Register gRPC server endpoint
  // Note: Make sure the gRPC server is running properly and accessible
  mux := runtime.NewServeMux()
  opts := []grpc.DialOption{grpc.WithTransportCredentials(insecure.NewCredentials())}
  err := pb.RegisterBookstoreHandlerFromEndpoint(ctx, mux,  *grpcServerEndpoint, opts)
  if err != nil {
    return err
  }

  // Start HTTP server (and proxy calls to gRPC server endpoint)
  return http.ListenAndServe(":8081", mux)
}

func main() {
  flag.Parse()
  defer glog.Flush()

  if err := run(); err != nil {
    glog.Fatal(err)
  }
}

执行如下命令，启动gateway服务， 监听8081端口:

go mod tidy
go run main.go

查看效果:

$ curl 127.0.0.1:8081/v1/shelves
{"shelves":[{"id":"1", "theme":"Fiction"}, {"id":"2", "theme":"Fantasy"}]}

Envoy

安装envoy: https://www.envoyproxy.io/docs/envoy/latest/start/install#tools-images

生成descriptor file

mkdir envoy
cd envoy
cp ../http_bookstore.proto ./

# 因为http_bookstore.proto依赖googleapi的proto file
git clone https://github.com/googleapis/googleapis.git

# 生成 descriptor file
protoc -o api_descriptor.pb -I ./googleapis/ -I ./  --include_imports  http_bookstore.proto

创建conig.yaml, 添加如下内容:

static_resources:
  listeners:
  - name: listener1
    address:
      socket_address: {address: 0.0.0.0, port_value: 8080}
    filter_chains:
    - filters:
      - name: envoy.filters.network.http_connection_manager
        typed_config:
          "@type": type.googleapis.com/envoy.extensions.filters.network.http_connection_manager.v3.HttpConnectionManager
          stat_prefix: grpc_json
          codec_type: AUTO
          route_config:
            name: local_route
            virtual_hosts:
            - name: local_service
              domains: ["*"]
              routes:
              # NOTE: by default, matching happens based on the gRPC route, and not on the incoming request path.
              # Reference: https://envoyproxy.io/docs/envoy/latest/configuration/http/http_filters/grpc_json_transcoder_filter#route-configs-for-transcoded-requests
              - match: {prefix: "/endpoints.examples.bookstore.Bookstore"}
                route: {cluster: grpc, timeout: 60s}
          http_filters:
          - name: envoy.filters.http.grpc_json_transcoder
            typed_config:
              "@type": type.googleapis.com/envoy.extensions.filters.http.grpc_json_transcoder.v3.GrpcJsonTranscoder
              proto_descriptor: "/home/gpx/demorpc/envoy/api_descriptor.pb"
              services: ["endpoints.examples.bookstore.Bookstore"]
              print_options:
                add_whitespace: true
                always_print_primitive_fields: true
                always_print_enums_as_ints: false
                preserve_proto_field_names: false
          - name: envoy.filters.http.router

  clusters:
  - name: grpc
    type: STATIC
    lb_policy: ROUND_ROBIN
    connect_timeout: 2s
    dns_lookup_family: V4_ONLY
    typed_extension_protocol_options:
      envoy.extensions.upstreams.http.v3.HttpProtocolOptions:
        "@type": type.googleapis.com/envoy.extensions.upstreams.http.v3.HttpProtocolOptions
        explicit_http_config:
          http2_protocol_options: {}
    load_assignment:
      cluster_name: grpc
      endpoints:
      - lb_endpoints:
        - endpoint:
            address:
              socket_address:
                address: 127.0.0.1
                port_value: 8000

启动envoy

envoy -c config.yaml

验证结果:

$ curl 127.0.0.1:8080/v1/shelves
{
 "shelves": [
  {
   "id": "1",
   "theme": "Fiction"
  },
  {
   "id": "2",
   "theme": "Fantasy"
  }
 ]
}

istio

因为istio使用的sidecar也是envoy，所以也可以配置istio的sidecar做这样的转码

istio并为提供直接设置grpc转码的配置，但istio开了直接patch sidecar的envoy配置的扣子： EnvoyFilter，可以实现上面Envoy的配置

同样使用bookstore的例子
部署如下的deploy + service:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: bookstore
  labels:
    app: bookstore

spec:
  replicas: 1
  selector:
    matchLabels:
      app: bookstore
  template:
    metadata:
      labels:
        app: bookstore
    spec:
      containers:
      - name: bookstore
        image: dockerhub.nie.netease.com/xiebaogong/bookstore:v1
        args:
        - "--port"
        - "9090"
        ports:
        - containerPort: 9090
---
apiVersion: v1
kind: Service
metadata:
  name: bookstore
  labels:
    app: bookstore
    service: bookstore
spec:
  ports:
  - port: 9091
    targetPort: 9090
    name: grpc
    appProtocol: grpc
  selector:
    app: bookstore

进入bookstore容器，验证grpc服务部署成功

# python bookstore_client.py  --host=bookstore --port=9091
ListShelves: shelves {
  id: 1
  theme: "Fiction"
}
shelves {
  id: 2
  theme: "Fantasy"
}

配置EnvoyFilter

apiVersion: networking.istio.io/v1alpha3
kind: EnvoyFilter
metadata:
  name: bookstore-grpc-json
spec:
  workloadSelector:
    labels:
      app: bookstore
  configPatches:
  - applyTo: HTTP_FILTER
    match:
      context: SIDECAR_INBOUND
      listener:
        portNumber: 9090
        filterChain:
          filter:
            name: "envoy.filters.network.http_connection_manager"
            subFilter:
              name: "envoy.filters.http.router"
    patch:
      operation: INSERT_BEFORE
      value:
        name: envoy.filters.http.grpc_json_transcoder
        typed_config:
          "@type": type.googleapis.com/envoy.extensions.filters.http.grpc_json_transcoder.v3.GrpcJsonTranscoder
          proto_descriptor_bin: Ctd4ChVnb29nbGUvYXBpL2h0dH....
          print_options:
            add_whitespace: true
            always_print_primitive_fields: true
            always_print_enums_as_ints: false
            preserve_proto_field_names: false

留意上面的 proto_descriptor_bin配置， 取值内容是上面Envoy demo中的api_descriptor.pb转成base64，copy上去
使用 kubectl apply -f filter.yaml 部署上述配置之后，可在k8s的任意容器中验证结果:

$ curl bookstore:9091/v1/shelves
{
 "shelves": [
  {
   "id": "1",
   "theme": "Fiction"
  },
  {
   "id": "2",
   "theme": "Fantasy"
  }
 ]

因为EnvoyFilter直接暴露了Envoy的配置，在不通版本下不太一样，上面的测试是在istio 1.11上完成的，在istio 1.4下配置不通，如下：

apiVersion: networking.istio.io/v1alpha3
kind: EnvoyFilter
metadata:
  name: bookstore-grpc-json
  namespace: cbg-xie1
spec:
  workloadSelector:
    labels:
      app: bookstore
  configPatches:
  - applyTo: HTTP_FILTER
    match:
      context: SIDECAR_INBOUND
      listener:
        portNumber: 9390
        filterChain:
          filter:
            name: "envoy.http_connection_manager"
            subFilter:
              name: "envoy.router"
    patch:
      operation: INSERT_BEFORE
      value:
        name: envoy.grpc_json_transcoder
        config:
          auto_mapping: true
          proto_descriptor_bin: Ctd4ChVnb29nbGUvYXB...
          services: ["endpoints.examples.bookstore.Bookstore"]
          print_options:
            add_whitespace: true
            always_print_primitive_fields: true
            always_print_enums_as_ints: false
            preserve_proto_field_names: false

改配置部署成功，并通过istioctl proxy-config listener 查看，配置确实写入了， 但是访问并不生效，怀疑是bug

总结

gRPC Gateway, Envoy, istio配置这3种方式对比下来各有优劣

gRPC Gateway需要引入go，对于非go语言，引入go，流程有些复杂，不太友好
Envoy 再引入这么一个专门的proxy，比较重，也可以考虑和业务容器分开单独部署
istio配置，不同版本的兼容性不太好

最后编辑于 ：
©著作权归作者所有,转载或内容合作请联系作者
平台声明：文章内容（如有图片或视频亦包括在内）由作者上传并发布，文章内容仅代表作者本人观点，简书系信息发布平台，仅提供信息存储服务。

推荐阅读更多精彩内容