gistfile1.txt

ubuntu@ubuntu:~$ systemd-analyze critical-chain cloud-init-local.service
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.

cloud-init-local.service +1min 27.432s
└─systemd-remount-fs.service @35.183s +1.631s
  └─systemd-fsck-root.service @29.899s +3.711s
    └─systemd-journald.socket @26.502s
      └─system.slice @26.076s
        └─-.slice @26.073s

ubuntu@ubuntu:~$ systemd-analyze blame
3min 35.206s snapd.seeded.service
1min 27.432s cloud-init-local.service
     49.577s dev-vda1.device
     48.939s cloud-init.service
     36.818s snapd.service
     34.327s apport.service
     30.878s pollinate.service
     26.549s cloud-config.service
     22.788s rsyslog.service
     20.467s cloud-final.service
     20.259s systemd-udev-trigger.service
     16.607s systemd-resolved.service
     14.968s snap.lxd.activate.service
     12.838s dev-loop0.device
     12.520s dev-loop1.device
     11.969s keyboard-setup.service
     11.863s dev-loop2.device
     11.539s snapd.apparmor.service
     10.061s systemd-logind.service
      9.491s apparmor.service
      8.947s systemd-timesyncd.service
      7.936s update-notifier-download.service
      6.360s multipathd.service

Long Calls (manual analysis):
-----------------------------
12s - netplan generate
6s  - blkid
<6s - loading & parsing yaml
1s  - closing stdin (or whatever comes after that that is not logged)
.5s - mount
.5s - umount

The question is:

Why does cloud-init think it only took 40s, but systemd thinks it took 1m 30s?

2023-04-06 19:10:59,215 - util.py[DEBUG]: cloud-init mode 'init' took 40.976 seconds (40.96)
2023-04-06 19:10:59,219 - handlers.py[DEBUG]: finish: init-local: SUCCESS: searching for local datasources
2023-04-06 19:11:17,505 - util.py[DEBUG]: Cloud-init v. 23.1.1-0ubuntu2 running 'init' at Thu, 06 Apr 2023 19:11:16 +0000. Up 228.80 seconds.