zifeihan commented on a change in pull request #5718:
URL: https://github.com/apache/skywalking/pull/5718#discussion_r511801780
##########
File path:
oap-server/server-fetcher-plugin/kafka-fetcher-plugin/src/main/java/org/apache/skywalking/oap/server/analyzer/agent/kafka/KafkaFetcherHandlerRegister.java
##########
@@ -132,16 +157,21 @@ public void start() {
@Override
public void run() {
while (true) {
- ConsumerRecords<String, Bytes> consumerRecords =
consumer.poll(Duration.ofMillis(500L));
- if (!consumerRecords.isEmpty()) {
- Iterator<ConsumerRecord<String, Bytes>> iterator =
consumerRecords.iterator();
- while (iterator.hasNext()) {
- ConsumerRecord<String, Bytes> record = iterator.next();
- handlerMap.get(record.topic()).handle(record);
+ try {
Review comment:
@dmsolr hello, kafka consumer poll data is fast, and each partition can
only have one thread to consume. In our stress testing process, data pulled by
Kafka is not a performance bottleneck, but trace data analysis is the
performance bottleneck. So I think it's better to pull the data as it is.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
