Golang微服务：Micro限流、熔断

Wrapper

Wrapper提供了一种包装机制，使得在执行某方法前先执行Wrapper，优点Filter的意思；因此可以在客户端和服务器做很多功能：熔断限流、Filter、Auth等。
client代码如下：调用greeter.Hello时先执行logWrap.Call方法，再调用RPC请求。

// log wrapper logs every time a request is made
type logWrapper struct {
	client.Client
}

func (l *logWrapper) Call(ctx context.Context, req client.Request, rsp interface{}, opts ...client.CallOption) error {
	fmt.Printf("[wrapper] client request service: %s method: %s
", req.Service(), req.Method())
	return l.Client.Call(ctx, req, rsp)
}

// Implements client.Wrapper as logWrapper
func logWrap(c client.Client) client.Client {
	return &logWrapper{c}
}

func main() {
	service := micro.NewService(
		micro.Name("greeter.client"),
		micro.Registry(mdns.NewRegistry()),
		// wrap the client
		micro.WrapClient(logWrap),
	)

	service.Init()

	greeter := proto.NewGreeterService("greeter", service.Client())

	rsp, err := greeter.Hello(context.TODO(), &proto.HelloRequest{Name: "John"})
	if err != nil {
		fmt.Println(err)
		return
	}

	fmt.Println(rsp.Greeting)

}

server代码如下:当RPC调用进来时先执行logWrapper，再执行Hello

type Greeter struct{}

func (g *Greeter) Hello(ctx context.Context, req *proto.HelloRequest, rsp *proto.HelloResponse) error {
	rsp.Greeting = "Hello " + req.Name
	log.Println("in Hello")
	return nil
}

// logWrapper is a handler wrapper
func logWrapper(fn server.HandlerFunc) server.HandlerFunc {
	return func(ctx context.Context, req server.Request, rsp interface{}) error {
		log.Printf("[wrapper] server request: %v", req.Method())
		err := fn(ctx, req, rsp)
		return err
	}
}

func main() {
	service := micro.NewService(
		micro.Name("greeter"),
		// wrap the handler
		micro.WrapHandler(logWrapper),
		micro.Registry(mdns.NewRegistry()),
	)

	service.Init()

	proto.RegisterGreeterHandler(service.Server(), new(Greeter))

	if err := service.Run(); err != nil {
		fmt.Println(err)
	}
}

熔断

Micro提供了两种实现，gobreaker和hystrix，熔断是在客户端实现。先看看 hystrix：

var (
	// DefaultTimeout is how long to wait for command to complete, in milliseconds
	DefaultTimeout = 1000
	// DefaultMaxConcurrent is how many commands of the same type can run at the same time
	DefaultMaxConcurrent = 10
	// DefaultVolumeThreshold is the minimum number of requests needed before a circuit can be tripped due to health
	DefaultVolumeThreshold = 20
	// DefaultSleepWindow is how long, in milliseconds, to wait after a circuit opens before testing for recovery
	DefaultSleepWindow = 5000
	// DefaultErrorPercentThreshold causes circuits to open once the rolling measure of errors exceeds this percent of requests
	DefaultErrorPercentThreshold = 50
	// DefaultLogger is the default logger that will be used in the Hystrix package. By default prints nothing.
	DefaultLogger = NoopLogger{}
)

type Settings struct {
	Timeout                time.Duration
	MaxConcurrentRequests  int
	RequestVolumeThreshold uint64
	SleepWindow            time.Duration
	ErrorPercentThreshold  int
}

hystrix会根据这5个参数（超时时间、并发请求数、请求量、空歇床、错误率）来选择合适的服务进行调度，目前是使用的 hystrix提供的默认参数，不支持自定义参数，示例:

func TestBreaker(t *testing.T) {
	// setup
	r := mock.NewRegistry()
	s := selector.NewSelector(selector.Registry(r))

	c := client.NewClient(
		// set the selector
		client.Selector(s),
		// add the breaker wrapper
		client.Wrap(NewClientWrapper()),
	)

	req := c.NewRequest("test.service", "Test.Method", map[string]string{
		"foo": "bar",
	}, client.WithContentType("application/json"))

	var rsp map[string]interface{}

	// Force to point of trip
	for i := 0; i < (hystrix.DefaultVolumeThreshold * 3); i++ {
		c.Call(context.TODO(), req, rsp)
	}

	err := c.Call(context.TODO(), req, rsp)
	if err == nil {
		t.Error("Expecting tripped breaker, got nil error")
	}

	if err.Error() != "hystrix: circuit open" {
		t.Errorf("Expecting tripped breaker, got %v", err)
	}
}

gobreaker方案与hystrix类似，可以自定义参数。

限流

ratelimit可以在客户端做，也可以在服务端做；micro提供了两种方案：juju/ratelimit、uber/ratelimit。

客户端实现:

func TestRateClientLimit(t *testing.T) {
	// setup
	r := mock.NewRegistry()
	s := selector.NewSelector(selector.Registry(r))

	testRates := []int{1, 10, 20, 100}

	for _, limit := range testRates {
		b := ratelimit.NewBucketWithRate(float64(limit), int64(limit))

		c := client.NewClient(
			// set the selector
			client.Selector(s),
			// add the breaker wrapper
			client.Wrap(NewClientWrapper(b, false)),//fasle=快速失败?
		)

		req := c.NewRequest(
			"test.service",
			"Test.Method",
			&TestRequest{},
			client.WithContentType("application/json"),
		)
		rsp := TestResponse{}

		for j := 0; j < limit; j++ {
			err := c.Call(context.TODO(), req, &rsp)
			e := errors.Parse(err.Error())
			if e.Code == 429 {
				t.Errorf("Unexpected rate limit error: %v", err)
			}
		}

		err := c.Call(context.TODO(), req, rsp)
		e := errors.Parse(err.Error())
		if e.Code != 429 {
			t.Errorf("Expected rate limit error, got: %v", err)
		}
	}
}

• NewBucketWithRate入参为速率（QPS）和容量（CAP），比如每秒5个请求，最大保持50个活动的请求
• NewClientWrapper第二个参数wait，指示当受到限流时是否等待，如果是false即快速失败，返回（429,too mant request）

服务端实现(以下代码包含了客户端测试代码):

func TestRateServerLimit(t *testing.T) {
	// setup
	r := mock.NewRegistry()
	s := selector.NewSelector(selector.Registry(r))

	testRates := []int{1, 10, 20}

	for _, limit := range testRates {
		b := ratelimit.NewBucketWithRate(float64(limit), int64(limit))
		c := client.NewClient(client.Selector(s))

		name := fmt.Sprintf("test.service.%d", limit)

		s := server.NewServer(
			server.Name(name),
			// add registry
			server.Registry(r),
			// add the breaker wrapper
			server.WrapHandler(NewHandlerWrapper(b, false)),
		)

		type Test struct {
			*testHandler
		}

		s.Handle(
			s.NewHandler(&Test{new(testHandler)}),
		)

		if err := s.Start(); err != nil {
			t.Fatalf("Unexpected error starting server: %v", err)
		}

		if err := s.Register(); err != nil {
			t.Fatalf("Unexpected error registering server: %v", err)
		}

		req := c.NewRequest(name, "Test.Method", &TestRequest{}, client.WithContentType("application/json"))
		rsp := TestResponse{}

		for j := 0; j < limit; j++ {
			if err := c.Call(context.TODO(), req, &rsp); err != nil {
				t.Fatalf("Unexpected request error: %v", err)
			}
		}

		err := c.Call(context.TODO(), req, &rsp)
		if err == nil {
			t.Fatalf("Expected rate limit error, got nil: rate %d, err %v", limit, err)
		}

		e := errors.Parse(err.Error())
		if e.Code != 429 {
			t.Fatalf("Expected rate limit error, got %v", err)
		}

		s.Deregister()
		s.Stop()

		// artificial test delay
		time.Sleep(time.Millisecond * 20)
	}
}

client.NewClient支持多个Wrapper，将熔断限流功能都添加上

	import "github.com/micro/go-plugins/wrapper/breaker/hystrix"
	import "github.com/micro/go-plugins/wrapper/ratelimiter/ratelimit"
	c := client.NewClient(
		client.Wrap(ratelimit.NewClientWrapper(b, false)),
		client.Wrap(hystrixNewClientWrapper()),
	)