hanahmily commented on issue #13522: URL: https://github.com/apache/skywalking/issues/13522#issuecomment-3343420423
The root cause is that **the liaison continues to send data to the data nodes without pause**, even when these nodes are experiencing memory pressure. This memory pressure arises from unflushed memory parts that the liaison has written. For a quick resolution, I will increase the number of working threads to speed up the flushing and merging process. According to the diagram, the CPU of the data node has the capacity to handle additional flushing activities. For long-term improvement, we need to implement a backpressure process to throttle the data flow from the liaison. @mrproliu you could work on reducing the traffic to prevent the data node from OOM. BTW, I didn't find the index data. Do you set the index rules for trace? -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
