From cb0c1aab803b99c581bd2be07e6f1739d0257623 Mon Sep 17 00:00:00 2001
From: abingcbc <abingcbc626@gmail.com>
Date: Sun, 25 Feb 2024 23:57:52 +0800
Subject: [PATCH] fix

---
 .../processor/extended/processor-rate-limit.md        |  4 ++--
 plugins/processor/ratelimit/processor_rate_limit.go   | 11 ++++-------
 plugins/processor/ratelimit/token_bucket.go           |  6 +++---
 3 files changed, 9 insertions(+), 12 deletions(-)

diff --git a/docs/cn/plugins/processor/extended/processor-rate-limit.md b/docs/cn/plugins/processor/extended/processor-rate-limit.md
index aa7d945135..f21f1fecb1 100644
--- a/docs/cn/plugins/processor/extended/processor-rate-limit.md
+++ b/docs/cn/plugins/processor/extended/processor-rate-limit.md
@@ -2,8 +2,8 @@
 
 ## 简介
 
-`processor_rate_limit processor`插件可以实现对日志的限速。在限定时间内，相同索引值下的日志只有低于等于限定速率数量的日志才会被采集，否则直接丢弃。
-例如，当索引字段为"ip"时，`'{"ip": "10.**.**.**", "method": "POST", "brower": "aliyun-sdk-java"}'`和`'{"ip": "10.**.**.**", "method": "GET", "brower": "aliyun-sdk-c++"}'`具有相同的索引值。因此，会累积计算`"ip": "10.**.**.**"`的日志数量。
+`processor_rate_limit processor`插件用于对日志进行限速处理，确保在设定的时间窗口内，具有相同索引值的日志条目的数量不超过预定的速率限制。若某索引值下的日志条目数量超出限定速率，则超额的日志条目将被丢弃不予采集。
+以"ip"字段作为索引的情况为例，考虑两条日志：`{"ip": "10.**.**.**", "method": "POST", "browser": "aliyun-sdk-java"}` 和`{"ip": "10.**.**.**", "method": "GET", "browser": "aliyun-sdk-c++"}`。这两条日志有相同的"ip"索引值（即 "10..."）。在此情形下，系统将对所有"ip"为"10..."的日志条目进行累计，确保其数量在限定时间窗口内不超过设定的速率限制。
 
 ## 版本
 
diff --git a/plugins/processor/ratelimit/processor_rate_limit.go b/plugins/processor/ratelimit/processor_rate_limit.go
index 1f5441a6e9..9bc928cfa1 100644
--- a/plugins/processor/ratelimit/processor_rate_limit.go
+++ b/plugins/processor/ratelimit/processor_rate_limit.go
@@ -59,10 +59,7 @@ func (p *ProcessorRateLimit) ProcessLogs(logArray []*protocol.Log) []*protocol.L
 	totalLen := len(logArray)
 	nextIdx := 0
 	for idx := 0; idx < totalLen; idx++ {
-		key, err := p.makeKey(logArray[idx])
-		if err != nil {
-			continue
-		}
+		key := p.makeKey(logArray[idx])
 		if p.Algorithm.IsAllowed(key) {
 			if idx != nextIdx {
 				logArray[nextIdx] = logArray[idx]
@@ -77,9 +74,9 @@ func (p *ProcessorRateLimit) ProcessLogs(logArray []*protocol.Log) []*protocol.L
 	return logArray
 }
 
-func (p *ProcessorRateLimit) makeKey(log *protocol.Log) (string, error) {
+func (p *ProcessorRateLimit) makeKey(log *protocol.Log) string {
 	if len(p.Fields) == 0 {
-		return "", nil
+		return ""
 	}
 
 	sort.Strings(p.Fields)
@@ -98,7 +95,7 @@ func (p *ProcessorRateLimit) makeKey(log *protocol.Log) (string, error) {
 		}
 	}
 
-	return strings.Join(values, "_"), nil
+	return strings.Join(values, "_")
 }
 
 func init() {
diff --git a/plugins/processor/ratelimit/token_bucket.go b/plugins/processor/ratelimit/token_bucket.go
index 1ffa21c70d..c643289f72 100644
--- a/plugins/processor/ratelimit/token_bucket.go
+++ b/plugins/processor/ratelimit/token_bucket.go
@@ -17,10 +17,10 @@ import (
 	"context"
 	"math"
 	"sync"
+	"sync/atomic"
 	"time"
 
 	"github.com/alibaba/ilogtail/pkg/logger"
-	"go.uber.org/atomic"
 )
 
 type bucket struct {
@@ -144,10 +144,10 @@ func (t *tokenBucket) runGC() {
 
 		// Add tokens to all buckets according to the rate limit
 		// and flag full buckets for deletion.
-		toDelete := make([]uint64, 0)
+		toDelete := make([]string, 0)
 		numBucketsBefore := 0
 		t.buckets.Range(func(k, v interface{}) bool {
-			key := k.(uint64)
+			key := k.(string)
 			b := v.(*bucket)
 
 			bucketFull := b.replenish(t.limit)