How to understand why the server becomes unavailable?

Question

linux

There is a dedicated server on hetzner, CentOS 6.6, Nginx and other standard web set.
From time to time, for no apparent reason, the server stops showing signs of life - all incoming requests on all available ports (80,443,22) go into a timeout, similar to the case of incorrectly configured iptables.
Nothing suspicious is observed in all kinds of logs - they simply break off. After rebooting from the hoster panel, everything is fine, until the next time. The problem has been repeating for almost a year approximately every 20-40 days, no systemicity has been noticed. The rest of the server does not cause any complaints.
I contacted technical support - I was refused with the wording "everything is ok with the hardware, look for a problem in your system."
Everything is ok with the system too, since a couple of months ago I reinstalled the system from scratch (for other reasons), the problem did not disappear.
How to determine what is causing this behavior?

Answer 1

what's in the logs, the system doesn't crash into a kernel panic, can it be accessed from the control panel or is it a system restart like when a button is pressed?

Answer 2

As an option, some module falls off. In general, turn on stronger information content in the logs (maybe even debug), after reboot, look at them (syslog, dmesg)

Answer 3

There was a similar situation with the server.
Kernel update helped.

Answer 4

How do CPU and memory graphs behave? What software is on the server?