TIL about systemd-cgtop from here
when trying to debug tailscaled high CPU usage.
systemd-cgtop is interactive, but using:
systemd-cgtop -c
will sort the “control groups” by CPU usage.
Example output:
Control Group Tasks %CPU Memory Input/s Output/s
user.slice 63 1.1 401.0M - -
user.slice/user-1000.slice 63 1.1 401.0M - -
/ 153 1.0 413.1M - -
user.slice/user-1000.slice/session-1.scope 4 0.8 164.3M - -
user.slice/user-1000.slice/session-c1.scope 50 0.4 226.7M - -
system.slice 34 0.3 174.3M - -
system.slice/tailscaled.service 10 0.1 39.4M - -
system.slice/systemd-oomd.service 1 0.1 1.4M - -
system.slice/systemd-journald.service 1 0.1 45.9M - -
dev-hugepages.mount - - 368.0K - -
dev-mqueue.mount - - 8.0K - -
init.scope 1 - 5.6M - -
sys-fs-fuse-connections.mount - - 8.0K - -
sys-kernel-config.mount - - 4.0K - -
sys-kernel-debug.mount - - 4.0K - -
system.slice/dbus.service 1 - 2.8M - -
system.slice/dev-zram0.swap - - 356.0K - -
system.slice/nix-daemon.service 2 - 4.9M - -
system.slice/nscd.service 11 - 3.5M - -
system.slice/sshd.service 1 - 5.0M - -
system.slice/system-getty.slice 1 - 280.0K - -
system.slice/system-getty.slice/getty@tty1.service 1 - 276.0K - -
system.slice/system-modprobe.slice - - 724.0K - -
system.slice/systemd-logind.service 1 - 2.0M - -
system.slice/systemd-networkd.service 1 - 3.1M - -
system.slice/systemd-resolved.service 1 - 9.6M - -
system.slice/systemd-timesyncd.service 2 - 1.4M - -
system.slice/systemd-udevd.service 1 - 19.5M - -
user.slice/user-1000.slice/session-c3.scope 7 - 2.3M - -
user.slice/user-1000.slice/user@1000.service 2 - 6.7M - -
Man page: https://www.freedesktop.org/software/systemd/man/systemd-cgtop.html
Using top or htop works but I have no idea how to filter out regular processes from the list.
Post 19 of #100DaysToOffload.