Server downtime due to frequent comments from an IP address

In the morning, I suddenly received a notification about the downtime of dnspod ).

 

The server reached 85% at AM. Uptime indicates that the CPU usage reaches 35. Don't go down.

 

Based on previous experience, it should trigger a CPU-consuming processing, restart PHP-CGI to restore immediately, and then view the log.

 

OK immediately after restart.

 

View the log and call up the log at that time point one by one, focusing on the response time. The normal processing time should be within 1 second, and many logs that have been found for more than 10 seconds are traced back slowly. The records of the first exception are found, and multiple logs are submitted for comment.

 

We found a series of high-frequency browsing of concurrent logs from the same IP address (occasionally there are other IP addresses in the same IP address segment, and checked from the United States ). Because our review is the third service used, it takes a lot of time.

How can I solve the problem?

Do you want to stop using the review function or simply add IP addresses to the blacklist?

 

Some APEC countries are still holding meetings during this period of time, so we should be cautious about the review. We should first add this IP segment to the blacklist.

Server downtime due to frequent comments from an IP address