0Pricing
NestJS Enterprise Backend APIs · 课时

实现并调用 gRPC 方法

连接 @GrpcMethod 处理器和 ClientGrpc 代理，处理一元请求/响应调用。

实现并调用 gRPC 方法 是 CoddyKit 上的免费 NestJS Enterprise Backend APIs 课时。 这是第 2 节课，共 4 节。 你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。 这是 NestJS Enterprise Backend APIs 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 NestJS Enterprise Backend APIs 课程共包含 4 节课。

本课时的部分内容尚未翻译，以英文显示。

Unary gRPC in NestJS

gRPC services in NestJS are defined by a .proto contract and implemented as ordinary providers decorated with gRPC handler metadata. The most common interaction is the unary call: the client sends a single request message and receives a single response message.

  • The server exposes handlers via @GrpcMethod (or @GrpcStreamMethod for streams).
  • The client obtains a typed proxy through ClientGrpc.getService() and calls methods that return Observables.

In this lesson we wire both ends of a unary request/response flow for an enterprise-style UsersService.

The .proto contract

Everything starts with the service contract. The package name and the service / rpc names are what NestJS uses to bind handlers and to resolve the client proxy.

  • package users maps to the transport option package: 'users'.
  • FindOne is the RPC NestJS will route to a matching handler.
syntax = "proto3";

package users;

service UsersService {
  rpc FindOne (UserById) returns (User) {}
}

message UserById {
  int32 id = 1;
}

message User {
  int32 id = 1;
  string name = 2;
  string email = 3;
}

Configuring the gRPC microservice

On the server, you start a microservice with the GRPC transport. The two critical options are package (must match the .proto package) and protoPath (where the contract lives).

  • url sets the bind address; default is localhost:5000.
  • You can pass an array of packages and proto paths for multi-service apps.
import { NestFactory } from '@nestjs/core';
import { Transport, MicroserviceOptions } from '@nestjs/microservices';
import { join } from 'path';
import { AppModule } from './app.module';

async function bootstrap() {
  const app = await NestFactory.createMicroservice<MicroserviceOptions>(
    AppModule,
    {
      transport: Transport.GRPC,
      options: {
        package: 'users',
        protoPath: join(__dirname, 'users.proto'),
        url: '0.0.0.0:5000',
      },
    },
  );
  await app.listen();
}
bootstrap();

Implementing a @GrpcMethod handler

A controller method becomes a unary RPC handler when you annotate it with @GrpcMethod. The decorator takes the service name and optionally the method name.

  • If you omit the method name, NestJS uses the PascalCase of the handler method name (so a method named findOne binds to FindOne).
  • The first argument is the deserialized request message; you simply return the response object (or a Promise/Observable of it).
import { Controller } from '@nestjs/common';
import { GrpcMethod } from '@nestjs/microservices';

interface UserById { id: number; }
interface User { id: number; name: string; email: string; }

@Controller()
export class UsersController {
  private readonly users: User[] = [
    { id: 1, name: 'Ada', email: 'ada@corp.io' },
    { id: 2, name: 'Linus', email: 'linus@corp.io' },
  ];

  @GrpcMethod('UsersService', 'FindOne')
  findOne(data: UserById): User {
    return this.users.find((u) => u.id === data.id);
  }
}

Method name resolution rules

Binding depends entirely on naming. Get this wrong and the call fails at runtime with an UNIMPLEMENTED error.

  • Service name in @GrpcMethod('UsersService') must match the service in the .proto.
  • Method name: explicit second arg wins; otherwise the handler method name is capitalized to PascalCase.

For clarity in enterprise code, prefer passing both arguments explicitly so renames of the TypeScript method cannot silently break the wire contract.

Registering the client with ClientsModule

To consume a gRPC service, register a client in the consuming module. Each entry gets a name (an injection token) and the same transport options as the server.

  • package and protoPath point at the same contract the server uses.
  • url targets the server's bind address.
import { Module } from '@nestjs/common';
import { ClientsModule, Transport } from '@nestjs/microservices';
import { join } from 'path';
import { ApiController } from './api.controller';

@Module({
  imports: [
    ClientsModule.register([
      {
        name: 'USERS_PACKAGE',
        transport: Transport.GRPC,
        options: {
          package: 'users',
          protoPath: join(__dirname, 'users.proto'),
          url: 'users-svc:5000',
        },
      },
    ]),
  ],
  controllers: [ApiController],
})
export class ApiModule {}

Getting the typed proxy with ClientGrpc

The injected client is a ClientGrpc instance, not the service itself. You must call getService() once the module is ready to obtain the strongly-typed proxy.

  • Resolve the proxy in onModuleInit so it exists before any request is handled.
  • The generic argument (UsersServiceClient) gives you full type safety on method calls.
import { Controller, Inject, OnModuleInit } from '@nestjs/common';
import { ClientGrpc } from '@nestjs/microservices';
import { Observable } from 'rxjs';

interface User { id: number; name: string; email: string; }
interface UsersServiceClient {
  findOne(data: { id: number }): Observable<User>;
}

@Controller('users')
export class ApiController implements OnModuleInit {
  private usersService: UsersServiceClient;

  constructor(@Inject('USERS_PACKAGE') private client: ClientGrpc) {}

  onModuleInit() {
    this.usersService = this.client.getService<UsersServiceClient>('UsersService');
  }
}

Calling a unary method returns an Observable

The proxy's methods are generated from the .proto and each returns an RxJS Observable, even for unary calls. NestJS can return the Observable directly from an HTTP handler, or you can convert it to a Promise.

  • Use firstValueFrom from RxJS when you need async/await ergonomics.
  • The proxy method name is the camelCase of the RPC (FindOne → findOne).
import { Controller, Get, Param, Inject, OnModuleInit } from '@nestjs/common';
import { ClientGrpc } from '@nestjs/microservices';
import { firstValueFrom, Observable } from 'rxjs';

interface User { id: number; name: string; email: string; }
interface UsersServiceClient {
  findOne(data: { id: number }): Observable<User>;
}

@Controller('users')
export class ApiController implements OnModuleInit {
  private usersService: UsersServiceClient;
  constructor(@Inject('USERS_PACKAGE') private client: ClientGrpc) {}

  onModuleInit() {
    this.usersService = this.client.getService<UsersServiceClient>('UsersService');
  }

  @Get(':id')
  async getUser(@Param('id') id: string): Promise<User> {
    return firstValueFrom(this.usersService.findOne({ id: Number(id) }));
  }
}

Why camelCase vs PascalCase matters

There are two distinct name transformations and mixing them up is a frequent bug source:

  • Server side: @GrpcMethod binds to the PascalCase RPC name (FindOne).
  • Client side: the proxy exposes the camelCase method (findOne) regardless of how the RPC is spelled in the proto.

So you call usersService.findOne(...) on the client even though the RPC and the server handler reference FindOne.

Mapping the Observable pipeline

Because unary calls return Observables, you can compose them with RxJS operators before exposing the result. This is idiomatic when you need to reshape or enrich the gRPC response.

  • This pure RxJS example mirrors the shape of a gRPC unary response without needing a running server, so it can execute standalone.
import { of, firstValueFrom } from 'rxjs';
import { map } from 'rxjs/operators';

interface User { id: number; name: string; email: string; }

// Simulates this.usersService.findOne({ id: 1 })
function findOne(data: { id: number }) {
  const row: User = { id: data.id, name: 'Ada', email: 'ada@corp.io' };
  return of(row);
}

async function main() {
  const dto = await firstValueFrom(
    findOne({ id: 1 }).pipe(
      map((u) => ({ userId: u.id, label: `${u.name} <${u.email}>` })),
    ),
  );
  console.log(JSON.stringify(dto));
}

main();

Handling errors with gRPC status codes

In enterprise services you should surface domain failures as proper gRPC statuses, not generic exceptions. Throw an RpcException with a numeric code from @grpc/grpc-js so callers can react deterministically.

  • status.NOT_FOUND (5) for a missing entity, status.INVALID_ARGUMENT (3) for bad input.
  • The client receives the status on the Observable's error channel.
import { Controller } from '@nestjs/common';
import { GrpcMethod, RpcException } from '@nestjs/microservices';
import { status } from '@grpc/grpc-js';

interface UserById { id: number; }
interface User { id: number; name: string; email: string; }

@Controller()
export class UsersController {
  private readonly users: User[] = [
    { id: 1, name: 'Ada', email: 'ada@corp.io' },
  ];

  @GrpcMethod('UsersService', 'FindOne')
  findOne(data: UserById): User {
    const found = this.users.find((u) => u.id === data.id);
    if (!found) {
      throw new RpcException({
        code: status.NOT_FOUND,
        message: `User ${data.id} not found`,
      });
    }
    return found;
  }
}

Quick Check

You implemented a server handler with @GrpcMethod('UsersService', 'FindOne'). On the consuming side you injected a ClientGrpc and called getService<UsersServiceClient>('UsersService'). Which method name do you invoke on the returned proxy to trigger this RPC?

Recap

You wired a complete unary gRPC flow in NestJS:

  • Defined the contract in a .proto with a package, service, and rpc.
  • Started a Transport.GRPC microservice and implemented the handler with @GrpcMethod('UsersService', 'FindOne'), returning the response object.
  • Registered a client via ClientsModule.register, resolved the typed proxy in onModuleInit with ClientGrpc.getService().
  • Called the camelCase proxy method, which returns an Observable — convertible with firstValueFrom.
  • Reported failures with RpcException and proper gRPC status codes.

Remember the naming split: server binds PascalCase, client calls camelCase.

常见问题解答

「实现并调用 gRPC 方法」课时是免费的吗？

是的 — 「实现并调用 gRPC 方法」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 NestJS Enterprise Backend APIs 课程的其余内容，请升级到 CoddyKit PRO。 NestJS Enterprise Backend APIs 课程共包含 4 节课。

「实现并调用 gRPC 方法」这节课中我会学到什么？

连接 @GrpcMethod 处理器和 ClientGrpc 代理，处理一元请求/响应调用。 你通过在浏览器中直接运行的动手代码来练习 NestJS Enterprise Backend APIs，全天候 AI 导师会在你学习这节课的过程中回答你的问题。

学习 NestJS Enterprise Backend APIs 需要有经验吗？

无需任何先前经验。CoddyKit 上的 NestJS Enterprise Backend APIs 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。 这是第 2 节课，共 4 节。

「实现并调用 gRPC 方法」课时需要多长时间？

大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。

我能在这节 NestJS Enterprise Backend APIs 课中编写并运行代码吗？

能。每节 NestJS Enterprise Backend APIs 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。

此课程中的所有课时

  1. 在 Protobuf 中定义服务与消息
  2. 实现并调用 gRPC 方法
  3. 流式 RPC 与背压
  4. 契约演进与向后兼容
← 返回 NestJS Enterprise Backend APIs