fix: disable ratelimiter's slack

'Slack allows the limiter to accumulate "unspent" requests for future bursts of traffic.'

It can cause a large number of packets to be 'sent' suddenly after running for a while, resulting in a large number of packet losses.

This can be observed in the send/recv stats log.

internal/engine/resolver.go

@@ -69,7 +69,7 @@ func (e *Engine) resolver(wg *sync.WaitGroup) {
 	}()
 
 	go func() {
-		rl := ratelimit.New(conf.C.Rate)
+		rl := ratelimit.New(conf.C.Rate, ratelimit.WithoutSlack)
 		for t := range e.toResolver {
 			if _, ok := taskMap.Load(t.DomainName); ok {
 				continue