Step 1: Log in to the background server/monitoring platform to check whether the system resources have reached the upper limit, such as: CPU, memory, disk, I/O, network bandwidth, etc. If these problems are encountered, solve these problems one by one:
If it is a CPU problem, you need to check the process with a relatively high CPU ratio, and then use the jstack command to generate the stack information of the process to see if frequent Full GC occurs. If so, you need to check the memory snapshot and analyze the memory situation. (You can use the built-in java or third-party tools); if the disk space is full, clear the disk in time; if the bandwidth is full, contact a network engineer to solve it. If none of the above problems exist, go to step 2.
Step 2: Check whether the thread pool configuration of the application server (Jboss/Tomcat) is reasonable, and check whether the queuing of requests is serious. If it is serious, you need to reset a reasonable thread pool. Similarly, check whether the connection pool setting of the database is reasonable, increase the connection pool setting, and check whether there is slow SQL. If there is slow SQL, optimize it (the optimization plan is to check the execution plan, set a reasonable index, etc.).
Step 3: Check the call chain of the slow-access service, and check whether the response time of each step in the call chain is reasonable. If it is not reasonable, contact the person in charge of the relevant system to investigate and solve it.
Step 4: Check the request log of the web server to see if there is a Doss attack. If there is a Doss attack, add the attacker's IP to the firewall's blacklist.