In recent pinpoint monitoring, I have observed that occasionally, individual threads will run for a significantly long time, even exceeding the proxy read timeout set on nginx, which is 60 seconds. Despite returning a 504 error due to the timeout, these threads still continue to execute. I am wondering if it is possible to terminate these threads or identify the cause of their prolonged execution time. Upon investigation, I found that the slow execution is mainly caused by prolonged JDBC queries.