I was told by my server admin that the server CPU reached 100% yesterday evening, today I checked the sa file of yeserday under /var/log/sa/, and I found there was no data suddenly in the file when CPU issue happened and sar data resumed after this server was reboot

May I get to know why no sar data updated for more than 1 hour until the server reboot? does Linus stop updating sar file when CPU is high? Thanks.

sar data in the file:

12:40:02 PM all 7.27 0.00 1.98 0.04 0.00 90.71
12:50:01 PM all 7.61 0.00 2.00 0.04 0.00 90.35
01:00:01 PM all 7.39 0.00 1.72 0.03 0.00 90.86
01:10:01 PM all 8.11 0.00 2.51 0.03 0.00 89.35
01:20:01 PM all 7.97 0.00 2.38 0.04 0.00 89.61
01:30:01 PM all 7.44 0.00 1.89 0.04 0.00 90.63
01:40:01 PM all 8.39 0.00 2.63 0.09 0.00 88.89
01:50:01 PM all 7.81 0.00 2.17 0.07 0.00 89.95
02:00:01 PM all 7.74 0.00 2.23 0.04 0.00 89.99
Average: all 7.36 0.00 1.80 0.04 0.00 90.81

03:10:26 PM LINUX RESTART

03:20:01 PM CPU %user %nice %system %iowait %steal %idle
03:30:01 PM all 2.36 0.00 1.91 0.09 0.00 95.64
03:40:01 PM all 1.91 0.00 1.20 0.14 0.00 96.75
03:50:01 PM all 1.82 0.00 1.08 0.04 0.00 97.05
04:00:01 PM all 1.85 0.00 1.15 3.18 0.00 93.82
04:10:01 PM all 1.85 0.00 1.02 7.17 0.00 89.96
04:20:02 PM all 1.56 0.00 0.77 0.03 0.00 97.64

Most likely because whatever made your CPU go to 100% prevent sar from being able to run.

When you say you rebooted do you mean you did a warm reboot or did you have to power off the system because a warm boot couldn't be done?

It happens on occasion that so many resources are used the system will be totally unresponsive requiring a hard boot (power cycle). In such occasions the same thing that prevents you from doing other tasks will prevent system jobs from running.

Another possibility is that the filesystem was full and rebooting cleaned up some temporary file that had filled it up.

Nothing in the output you put here shows any huge decrease of CPU idle by 2 PM. Presumably at next 2:10 PM sample whatever happened that sent your CPU to 100% had already occurred preventing that sample. Look at the rest of your sadd to see if you can determine other issues (e.g. memory or file table) growth that was increasing rapidly. The CPU going to a 100% may have been a symptom of the other issue.

Yes, I had to power off the system because the VM was not responsible. I also checked network, memory, fd and so on, everything was normal before the incident, it happened suddenly.
my server did not run on a physical machine but a VM managed by VMWare, maybe it is VM issue.
By the way, would you tell me how to check how much memory that Linux kernel is using? is it equal to a sum up of all output of /proc/slabinfo?