feat: 将 release 函数移动到 QuotaResponse 结构体中

Author: WTIFS

pkg/flow/quota/assist.go

@@ -241,7 +241,7 @@ func (f *FlowQuotaAssistant) GetQuota(commonRequest *data.CommonRateLimitRequest
 			Code: model.QuotaResultOk,
 			Info: Disabled,
 		}
-		return model.QuotaFutureWithResponse(resp, nil), nil
+		return model.QuotaFutureWithResponse(resp), nil
 	}
 	windows, err := f.lookupRateLimitWindow(commonRequest)
 	if err != nil {
@@ -253,33 +253,37 @@ func (f *FlowQuotaAssistant) GetQuota(commonRequest *data.CommonRateLimitRequest
 			Code: model.QuotaResultOk,
 			Info: RuleNotExists,
 		}
-		return model.QuotaFutureWithResponse(resp, nil), nil
+		return model.QuotaFutureWithResponse(resp), nil
 	}
 	var maxWaitMs int64 = 0
-	var releaseFuncs = make([]func(), 0, len(windows))
+	var releaseFuncs = make([]model.ReleaseFunc, 0, len(windows))
 	for _, window := range windows {
 		window.Init()
-		quotaResult, releaseFunc := window.AllocateQuotaWithRelease(commonRequest)
-		if releaseFunc != nil {
-			releaseFuncs = append(releaseFuncs, releaseFunc)
+		quotaResult := window.AllocateQuota(commonRequest)
+		if quotaResult == nil {
+			continue
+		}
+		for i := range quotaResult.ReleaseFuncs {
+			releaseFuncs = append(releaseFuncs, quotaResult.ReleaseFuncs[i])
 		}
 		// 触发限流，提前返回
 		if quotaResult.Code == model.QuotaResultLimited {
 			// 先释放资源
 			for i := range releaseFuncs {
-				releaseFuncs[i]()
+				releaseFuncs[i](0)
 			}
-			return model.QuotaFutureWithResponse(quotaResult, nil), nil
+			return model.QuotaFutureWithResponse(quotaResult), nil
 		}
 		// 未触发限流，记录令牌桶的最大排队时间
 		if quotaResult.WaitMs > maxWaitMs {
 			maxWaitMs = quotaResult.WaitMs
 		}
 	}
 	return model.QuotaFutureWithResponse(&model.QuotaResponse{
-		Code:   model.QuotaResultOk,
-		WaitMs: maxWaitMs,
-	}, releaseFuncs), nil
+		Code:         model.QuotaResultOk,
+		WaitMs:       maxWaitMs,
+		ReleaseFuncs: releaseFuncs,
+	}), nil
 }
 
 // lookupRateLimitWindow 计算限流窗口

pkg/flow/quota/window.go

@@ -608,15 +608,6 @@ func (r *RateLimitWindow) AllocateQuota(commonRequest *data.CommonRateLimitReque
 	return r.trafficShapingBucket.GetQuota(curTimeMs, commonRequest.Token)
 }
 
-// AllocateQuotaWithRelease 分配配额，并返回释放资源函数
-func (r *RateLimitWindow) AllocateQuotaWithRelease(commonRequest *data.CommonRateLimitRequest) (*model.QuotaResponse, func()) {
-	nowMilli := model.CurrentMillisecond()
-	atomic.StoreInt64(&r.lastAccessTimeMilli, nowMilli)
-	// 获取服务端时间
-	curTimeMs := r.toServerTimeMilli(nowMilli)
-	return r.trafficShapingBucket.GetQuotaWithRelease(curTimeMs, commonRequest.Token)
-}
-
 // GetLastAccessTimeMilli 获取最近访问时间
 func (r *RateLimitWindow) GetLastAccessTimeMilli() int64 {
 	return atomic.LoadInt64(&r.lastAccessTimeMilli)

pkg/model/quota.go

@@ -156,6 +156,8 @@ const (
 	QuotaResultLimited QuotaResultCode = -1
 )
 
+type ReleaseFunc func()
+
 // QuotaResponse 配额查询应答.
 type QuotaResponse struct {
 	// 配额分配的返回码
@@ -164,27 +166,27 @@ type QuotaResponse struct {
 	Info string
 	// 需要等待的时间段
 	WaitMs int64
+	// 释放资源函数
+	ReleaseFuncs []ReleaseFunc
 }
 
 // QuotaFutureImpl 异步获取配额的future.
 type QuotaFutureImpl struct {
-	resp         *QuotaResponse
-	deadlineCtx  context.Context
-	cancel       context.CancelFunc
-	releaseFuncs []func()
+	resp        *QuotaResponse
+	deadlineCtx context.Context
+	cancel      context.CancelFunc
 }
 
-func QuotaFutureWithResponse(resp *QuotaResponse, releaseFuncs []func()) *QuotaFutureImpl {
+func QuotaFutureWithResponse(resp *QuotaResponse) *QuotaFutureImpl {
 	var deadlineCtx context.Context
 	var cancel context.CancelFunc
 	if resp.WaitMs > 0 {
 		deadlineCtx, cancel = context.WithTimeout(context.Background(), time.Duration(resp.WaitMs)*time.Millisecond)
 	}
 	return &QuotaFutureImpl{
-		resp:         resp,
-		deadlineCtx:  deadlineCtx,
-		cancel:       cancel,
-		releaseFuncs: releaseFuncs,
+		resp:        resp,
+		deadlineCtx: deadlineCtx,
+		cancel:      cancel,
 	}
 }
 
@@ -209,10 +211,12 @@ func (q *QuotaFutureImpl) Get() *QuotaResponse {
 	return q.resp
 }
 
-// Release 释放资源，仅用于并发数限流的场景.
+// Release 释放资源，仅用于并发数限流/CPU限流场景
 func (q *QuotaFutureImpl) Release() {
-	for i := range q.releaseFuncs {
-		q.releaseFuncs[i]()
+	if q.resp != nil {
+		for i := range q.resp.ReleaseFuncs {
+			q.resp.ReleaseFuncs[i]()
+		}
 	}
 }
 

pkg/plugin/ratelimiter/ratelimiter.go

@@ -35,8 +35,6 @@ type ServiceRateLimiter interface {
 type QuotaBucket interface {
 	// GetQuota 在令牌桶/漏桶中进行单个配额的划扣，并返回本次分配的结果
 	GetQuota(curTimeMs int64, token uint32) *model.QuotaResponse
-	// GetQuotaWithRelease 判断限流结果，并返回配额释放函数（对并发数限流、CPU自适应限流有用）
-	GetQuotaWithRelease(curTimeMs int64, token uint32) (*model.QuotaResponse, func())
 	// Release 释放配额（仅对于并发数限流有用）
 	Release()
 	// OnRemoteUpdate 远程配额更新

plugin/ratelimiter/bbr/README.md

@@ -12,9 +12,7 @@ BBR 的源码实现可参考：
 
 
 # 插件设计
-本插件将 BBR 限流器适配成 `QuotaBucket` 接口（主要实现 `GetQuotaWithRelease` 判断限流方法），以及 `ServiceRateLimiter` 接口（实现 `InitQuota` 初始化方法）。
-
-由于 BBR 限流需要记录请求通过数、当前并发数、请求耗时，因此没有复用原来 `QuotaBucket` 接口中的 `GetQuota` 方法，而是新增了一个方法 `GetQuotaWithRelease`，该方法相比于 `GetQuota` 方法，返回参数中多了一个 `func()`，供业务方在业务逻辑处理完成后调用。
+本插件将 BBR 限流器适配成 `QuotaBucket` 接口（主要实现 `GetQuota` 判断限流方法），以及 `ServiceRateLimiter` 接口（实现 `InitQuota` 初始化方法）。
 
 由于 CPU 使用率指标为实例单机指标，因此 CPU 限流只适用于单机限流，不适用于分布式限流，未实现分布式限流器需要实现的接口。
 
@@ -29,7 +27,7 @@ bucket: 桶数，BBR 会把 window 分成多个 bucket，沿时间轴向前滑
 这三个入参，从 `apitraffic.Rule` 结构体中解析，直接使用了结构体中的 `MaxAmount`、`ValidDuration`、`Precision` 字段
 
 
-## 判断限流 GetQuotaWithRelease
+## 判断限流 GetQuota
 调用了 BBR 的 `Allow()` 方法
 
 其内部执行 `shouldDrop()` 方法，其执行流程如下：

plugin/ratelimiter/bbr/bucket.go

@@ -1,19 +1,32 @@
+/**
+ * Tencent is pleased to support the open source community by making polaris-go available.
+ *
+ * Copyright (C) 2019 THL A29 Limited, a Tencent company. All rights reserved.
+ *
+ * Licensed under the BSD 3-Clause License (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * https://opensource.org/licenses/BSD-3-Clause
+ *
+ * Unless required by applicable law or agreed to in writing, software distributed
+ * under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR
+ * CONDITIONS OF ANY KIND, either express or implied. See the License for the
+ * specific language governing permissions and limitations under the License.
+ */
+
 package bbr
 
 import (
-	"sort"
-
 	"github.com/polarismesh/polaris-go/pkg/model"
 	"github.com/polarismesh/polaris-go/pkg/plugin/ratelimiter"
 	"github.com/polarismesh/polaris-go/plugin/ratelimiter/bbr/core"
+	"sort"
 
 	apitraffic "github.com/polarismesh/specification/source/go/api/v1/traffic_manage"
 )
 
 var (
-	allowResp = &model.QuotaResponse{
-		Code: model.QuotaResultOk,
-	}
 	denyResp = &model.QuotaResponse{
 		Code: model.QuotaResultLimited,
 	}
@@ -26,20 +39,18 @@ type BBRQuotaBucket struct {
 
 // GetQuota 获取限额
 func (b *BBRQuotaBucket) GetQuota(_ int64, _ uint32) *model.QuotaResponse {
-	return nil
-}
-
-// GetQuotaWithRelease 判断是否限流，并返回释放资源函数
-func (b *BBRQuotaBucket) GetQuotaWithRelease(_ int64, _ uint32) (*model.QuotaResponse, func()) {
 	release, allow := b.BBR.Allow()
 	if allow {
-		return allowResp, release
+		return &model.QuotaResponse{
+			Code:         model.QuotaResultOk,
+			ReleaseFuncs: []model.ReleaseFunc{release},
+		}
 	}
-	return denyResp, nil
+	return denyResp
 }
 
-// Release 释放资源
-func (b *BBRQuotaBucket) Release() {
+// Release 释放配额（仅对于并发数限流有用）
+func (l *BBRQuotaBucket) Release() {
 
 }
 

plugin/ratelimiter/reject/bucket.go

@@ -31,11 +31,6 @@ func (q *QuotaBucketReject) GetQuota(curTimeMs int64, token uint32) *model.Quota
 	return q.bucket.Allocate(curTimeMs, token)
 }
 
-// GetQuotaWithRelease 在令牌桶/漏桶中进行单个配额的划扣，并返回本次分配的结果
-func (q *QuotaBucketReject) GetQuotaWithRelease(curTimeMs int64, token uint32) (*model.QuotaResponse, func()) {
-	return q.bucket.Allocate(curTimeMs, token), nil
-}
-
 // Release 释放配额（仅对于并发数限流有用）
 func (q *QuotaBucketReject) Release() {
 	q.bucket.Release()

plugin/ratelimiter/unirate/bucket_leaky.go

@@ -156,11 +156,6 @@ func (l *LeakyBucket) GetQuota(curTimeMs int64, token uint32) *model.QuotaRespon
 	return l.allocateQuota()
 }
 
-// GetQuotaWithRelease 在令牌桶/漏桶中进行单个配额的划扣，并返回本次分配的结果
-func (l *LeakyBucket) GetQuotaWithRelease(_ int64, _ uint32) (*model.QuotaResponse, func()) {
-	return l.allocateQuota(), nil
-}
-
 // Release 释放配额（仅对于并发数限流有用）
 func (l *LeakyBucket) Release() {