refactor: removed concurrent retry for failed log messages

We've commented the code for concurrent retries for failed log messages.
This was needed since we're flooding the NIC with too many requests and
it causes unexpected behaviours

Signed-off-by: jay-dee7 <jasdeepsingh.uppal@gmail.com>

telemetry/fluent-bit/fluent_bit.go

@@ -4,7 +4,6 @@ import (
 	"bytes"
 	"context"
 	"net/http"
-	"sync"
 	"time"
 
 	"github.com/containerish/OpenRegistry/config"
@@ -17,10 +16,8 @@ type (
 	}
 
 	fluentBit struct {
-		wg            sync.WaitGroup
 		client        *http.Client
 		retryMessages map[string]retryLogMsg
-		gate          chan struct{}
 		config        *config.RegistryConfig
 	}
 
@@ -39,8 +36,6 @@ func New(config *config.RegistryConfig) (FluentBit, error) {
 	fbClient := &fluentBit{
 		client:        httpClient,
 		config:        config,
-		wg:            sync.WaitGroup{},
-		gate:          make(chan struct{}, 5),
 		retryMessages: make(map[string]retryLogMsg),
 	}
 
@@ -101,19 +96,12 @@ func (fb *fluentBit) retry() {
 	// lets not do more than 5 req/second just to not flood our free instance of grafana cloud
 	for range ticker.C {
 		for id, logMsg := range fb.retryMessages {
-			fb.gate <- struct{}{}
-			fb.wg.Add(1)
-			go fb.retrier(logMsg.content, id)
+			fb.retrier(logMsg.content, id)
 		}
 	}
 }
 
 func (fb *fluentBit) retrier(logBytes []byte, id string) {
-	defer func() {
-		fb.wg.Done()
-		<-fb.gate
-	}()
-
 	// TODO - (@jay-dee7) what to do then? maybe have a different way to ship these logs? like via promtail?
 	if msg, ok := fb.retryMessages[id]; ok && msg.count > 3 {
 		delete(fb.retryMessages, id)