双Xeon E5-2683v4服务器上10个WordPress站点运行缓慢的问题排查求助

各位大佬好，我最近碰到个头疼的问题——我在一台双Xeon E5-2683v4的专用服务器上部署了10个WordPress站点，之前运行都挺稳定的，但最近速度突然暴跌，打开个站点居然要40秒！

先说明下基础情况：所有站点加起来总流量大概5k，不算高，但就是卡得离谱。我完全摸不着头绪，先跑了top命令查看系统状态，输出如下：

top - 04:05:52 up 123 days,  2:38,  1 user,  load average: 2.08, 2.11, 2.34

Tasks: 915 total,   5 running, 909 sleeping,   0 stopped,   1 zombie

%Cpu(s):  0.1 us,  0.6 sy,  5.3 ni, 94.0 id,  0.0 wa,  0.0 hi,  0.1 si,  0.0 st

MiB Mem : 128259.9 total,  40947.6 free,  14620.3 used,  72692.1 buff/cache

MiB Swap:   4096.0 total,   3068.7 free,   1027.2 used. 108475.6 avail Mem

PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND

2990823 user  30  10  518652 120496  24104 R 100.0   0.1   0:00.52 php-fpm

2990789 user  30  10  604468 239928 132724 R  94.1   0.2   0:06.29 php-fpm

2990816 user  30  10  569976 161748  89012 R  88.2   0.1   0:01.42 php-fpm

2990821 user  30  10  478804  82436  24500 R  88.2   0.1   0:00.96 php-fpm

2990614 user  30  10  610628 211532  98144 S  23.5   0.2   0:03.54 php-fpm

1981 mysql     20   0 5328392   2.6g  13348 S   5.9   2.1   6935:15 mysqld

2990825 root      20   0   55344   5344   3580 R   5.9   0.0   0:00.02 top

1 root      20   0  238960  11528   8288 S   0.0   0.0  96:29.69 systemd

2 root      20   0       0      0      0 S   0.0   0.0   0:10.52 kthreadd

3 root       0 -20       0      0      0 I   0.0   0.0   0:00.00 rcu_gp

4 root       0 -20       0      0      0 I   0.0   0.0   0:00.00 rcu_par_gp

从top结果能看到好几个php-fpm进程CPU直接拉满，但MySQL的CPU占用并不高，内存也还有大量富余（可用内存108G左右），交换区只用了一点，也不算异常。

之后我又跑了ps -aux进一步排查，输出如下：

user 2992194  0.0  0.0 188188 28144 ?        SN   04:09   0:00 /usr/local/apps/apache2/bin/httpd -k graceful

user 2992204 44.2  0.1 604516 239472 ?       SN   04:09   0:11 php-fpm: pool user-user

root     2992244  0.0  0.0      0     0 ?        I    04:09   0:00 [kworker/29:3-mm_percpu_wq]

root     2992245  0.0  0.0 189452  9700 ?        SN   04:09   0:00 php-fpm: pool index

root     2992248  0.0  0.0  44148  3428 ?        S    04:10   0:00 /usr/sbin/CROND -n

root     2992251  0.0  0.0  95596  6588 ?        S    04:10   0:00 /usr/sbin/CROND -n

root     2992252  0.0  0.0  12800  3112 ?        Ss   04:10   0:00 /bin/sh -c /usr/local/emps/bin/php /usr/local/webuzo/cro

root     2992254  0.3  0.0 196032 33104 ?        S    04:10   0:00 /usr/local/emps/bin/php /usr/local/webuzo/crons.php

user 2992257  0.0  0.0  12800  3036 ?        Ss   04:10   0:00 /bin/sh -c cd /home/user/public_html/v2 && php artis

user 2992258  2.1  0.0 500592 67244 ?        S    04:10   0:00 php artisan schedule:run

root     2992262  0.0  0.0      0     0 ?        I    04:10   0:00 [kworker/32:2-mm_percpu_wq]

root     2992406  0.4  0.0 195660 33244 ?        S    04:10   0:00 /usr/local/emps/bin/php -d auto_prepend_file=none -d aut

root     2992414  0.0  0.0  11996  2788 ?        S    04:10   0:00 sh -c /usr/local/emps/bin/php /usr/local/webuzo/cli.php

root     2992415  3.4  0.0 195836 33992 ?        D    04:10   0:00 /usr/local/emps/bin/php /usr/local/webuzo/cli.php --exim

user 2992420 79.7  0.1 607628 201180 ?       RN   04:10   0:11 php-fpm: pool user-user

user 2992425 89.4  0.1 534164 205828 ?       RN   04:10   0:11 php-fpm: pool user-user

root     2992441  0.2  0.0 121496  8432 ?        Ss   04:10   0:00 sshd: root [priv]

sshd     2992442  0.0  0.0  76724  2920 ?        S    04:10   0:00 sshd: root [net]

user 2992447 89.0  0.0 525656 128952 ?       RN   04:10   0:05 php-fpm: pool user-user

root     2992452  0.0  0.0  52748  3936 pts/1    R+   04:10   0:00 ps -aux

nobody   3275399  0.0  0.0 114496  5968 ?        SN   Jan07   0:05 nginx: cache manager process

nobody   3303798  0.0  0.0 114988  9492 ?        S    Jan07   0:06 nginx: cache manager process

named    3349620  0.0  0.6 5764496 799796 ?      Ssl   2023   8:07 /usr/sbin/named -u named -c /etc/named.conf

root     3420043  0.0  0.0 204652  2864 ?        S<sl  2023  11:39 /sbin/auditd

root     3456246  0.0  0.0  89556  8180 ?        Ss    2023   3:10 /usr/lib/systemd/systemd --user

root     3456248  0.0  0.0 285088  3484 ?        S     2023   0:00 (sd-pam)

root     3474540  0.0  0.0 169048  1200 ?        Ss    2023   1:54 gpg-agent --homedir /var/cache/dnf/remi-modular-6408ecca

root     3474597  0.0  0.0 169048  1304 ?        Ss    2023   1:55 gpg-agent --homedir /var/cache/dnf/remi-safe-ff04689114f

nobody   3832703  0.0  0.0 114660  5364 ?        SN   Jan09   0:04 nginx: cache manager process

root     3853593  0.0  0.0 166376 31076 ?        SNs  Jan09   0:41 /usr/local/apps/apache2/bin/httpd -k graceful

现在我完全搞不懂，流量不高但php-fpm进程CPU跑满，站点加载慢到离谱。有没有大佬能帮我分析下可能的原因？比如是不是某个WordPress插件出问题了？或者php-fpm配置不合理？或者MySQL有没有潜在的瓶颈？

备注：内容来源于stack exchange，提问作者Raden Bagus