Keywords: General - AWS - Technical issue - Other
bnsupport ID: f8707dc1-5cf1-9223-f98a-8632f310458e
Thanks in advance for help on this!
Summary: Received alert for surging database connections. Found multiple database connections are made and even with a bitnami restart, they seem to creep back up mechanically – fixed increasing rate each time. Some of the processes once killed revealed they were running for over an hour. There could have been some bot scanning activity that happened just before the behavior. Site is back online now but we’d like some additional guidelines because it is unclear if this will happen again.
Open Questions, Need Resolutions
What other settings can be added to prevent this behavior. For example, the killed processes were running for just over 54 minutes according to the php-fpm.log. They were running even after restart. Are there additional php-fpm or other conf settings that can be added to flush at restart?
It seems that once the spawning/surging php-fpm connections were finally killed, the 504 issues died off and the site was back up. It seems as if those php-fpm children were consuming all memory. Can you confirm or suggest a different explanation?
Based on the included access and error logs, can you assist further in this analysis and solution?
If/when this happened again, are there some other steps we can take to get your help as the issue is happening?
Details: I have a detailed, documented workflow that includes all the steps we took so far and screen shots. I ran the bitnami support tool and have a code. The access and error logs exceed 10mb - but I have a version with only today's entries.
The documentation and logs can be found here. Let me know once retrieved so I can delete: