Skip to content

Latest commit

 

History

History
91 lines (60 loc) · 3.56 KB

README.md

File metadata and controls

91 lines (60 loc) · 3.56 KB
Raw

SWAPI powered by gRPC Federation

Here is an example of building a simple service using gRPC Federation with SWAPI.

What is SWAPI ?

SWAPI is one of the public APIs. It frequently appears as an example for GraphQL and gRPC, so we are using it with gRPC Federation as well.

Background and Architecture

The service we are building this time with gRPC Federation is a BFF (Backend for Frontend) that consolidates the results of several pre-existing services ( film , person , planet , species , starship , vehicle ). Please consider these pre-existing services as in-house microservices in a real-world scenario. Each of them is developed and managed by a different team.

Each microservice exposes internal APIs using gRPC, with schemas managed by Protocol Buffers. Our goal is to build a gRPC service that aggregates the necessary information from each service and exposes it externally.

Architecture

How to use

The Protocol Buffers files for each microservice are located under the ./proto directory. Additionally, we will create the swapi.proto file for this project, which is already placed in ./proto/swapi.proto .

The swapi.proto file contains the API schema and the gRPC Federation options.

The following is an excerpt from it.

  • swapi.proto
// GetPersonReply message for SWAPI.
message GetPersonReply {
  // gRPC Federation Options.
  option (grpc.federation.message) = {
    def {
      name: "res"
      call {
        method: "swapi.person.PersonService/GetPerson"
        request { field: "id" by: "$.id" }
      }
    }
    ...
  };
}

In this demo, we will use Buf to generate code from the proto files utilizing gRPC Federation. First, you need to install the Buf CLI by running the following command:

make tools

If the installation is successful, the buf and ko and grpcurl binaries will be installed under the bin directory.

To generate code using Buf, execute the following command:

make generate

The code will be generated under the ./swapi/swapi directory.

Next, build the server and create the image. In this demo, it's necessary to start other microservices simultaneously, so the code for each is also provided. To build the code for all services and create images, execute the following command:

make build

Finally, start all services. Use docker compose for launching. The SWAPI service will be available on port 3000. ( See config file )

make up

Now, let's send a request to confirm the functionality. If you receive a response after sending a request like the following, the application is running successfully.

./bin/grpcurl -plaintext -d '{"id": 1}' localhost:3000 swapi.SWAPI/GetPerson

Additionally, we are using JAEGER to visualize OpenTelemetry traces. It is exposed on port 4000, so please try accessing it in your browser. If you can see a page like the following, it means it's successful.

  • http://localhost:4000

Jaeger UI

Conclusion

The team responsible for providing the SWAPI service needs to develop and maintain only the swapi.proto file and the ./cmd/swapi/main.go file. The swapi.proto file utilizes gRPC Federation. In the ./cmd/swapi/main.go file, additional tasks such as defining gRPC clients, which cannot be described solely with gRPC Federation, need to be addressed.

Please feel free to edit the swapi.proto file and explore the capabilities of gRPC Federation !!!