How does SOFABolt ensure machine performance issues?

[dajitui]

Your question

com.alipay.remoting.rpc.protocol.RpcRequestProcessor#process

As shown in the code on the left, "executor.execute(new ProcessTask(ctx, cmd))" inserts the result into the response future through the thread pool. In high concurrency scenarios, the thread pool may be overloaded, resulting in rejected requests. How to solve this problem? By adding more machine nodes or customizing thread pool monitoring?

Your scenes

Under high concurrency requests

Your advice

As a framework, custom thread pool has been implemented, which provides great convenience for developers, but also poses some hidden dangers. The thread pool also has corresponding bottlenecks. Are there any corresponding practical solutions to monitor the thread pool situation or properly configure the deployment of multiple nodes in certain ranges?

Environment

• SOFABolt version:1.6.6

[chuailiwu]

As you described, this thread pool can be customized and its metrics can be reported to your monitoring system.
Additionally, thread pool parameter configuration is strongly related to the business scenario and needs to be adjusted through stress testing analysis. I do not think there is a silver bullet solution.