偶尔在一篇博客看到，Linux 内核针对每个进程做了调度统计。内核文档如下：
/proc/<pid>/schedstat
----------------
schedstats also adds a new /proc/<pid>/schedstat file to include some of
the same information on a per-process level. There are three fields in
this file correlating for that process to:
1) time spent on the cpu
2) time spent waiting on a runqueue
3) # of timeslices run on this cpu
博客作者利用以上 schedstat 文件设计了一个工具：schedlat - 统计给定进程的调度延迟。示例如下：
# fio --ioengine=pvsync2 ... schedlat 采样 worker 进程
TIMESTAMP %CPU %LAT %SLP
2025-11-04 13:56:58 41.8 1.7 56.4
2025-11-04 13:56:59 40.8 1.7 57.5
2025-11-04 13:57:00 41.9 1.7 56.4
2025-11-04 13:57:01 41.9 1.6 56.5
2025-11-04 13:57:02 40.5 1.5 58.0
2025-11-04 13:57:03 39.0 1.7 59.3
...
# fio --ioengine=libaio ... schedlat 采样 worker 进程
TIMESTAMP %CPU %LAT %SLP
2025-11-04 13:57:34 87.0 4.5 8.5
2025-11-04 13:57:35 87.0 4.8 8.2
2025-11-04 13:57:36 86.7 5.1 8.1
2025-11-04 13:57:37 86.8 4.8 8.4
2025-11-04 13:57:38 86.8 4.6 8.5
2025-11-04 13:57:39 86.7 5.2 8.2
...
挺方便。