Pm2 changes cwd on hard reboot

What's going wrong?

I have only one node js app daemonized on server which i started with apps.json file
everything works fine until server shuts down
After that pm2 changes cwd value to user home folder ("/home/admin")
last time this happened when i forgot to pay server bill and my provider shut it down
I belive this happens on hard reboot

How could we reproduce this issue?

start node js app with apps.json file
pm2 start apps.json
pm2 startup
pm2 save
and hard reboot server

Supporting information

apps.json file

{
  "apps": [
    {
      "name": "sasa_front",
      "cwd": "/home/admin/apps/sasa_front",
      "script": "npm",
      "args": "start",
    }
  ]
}

--- Daemon logs --------------------------------------------
PM2        | 2022-07-23T10:23:58: PM2 log: App [sasa_front:0] starting in -fork mode-
PM2        | 2022-07-23T10:23:58: PM2 log: App [sasa_front:0] online
PM2        | 2022-07-23T10:23:59: PM2 log: App [sasa_front:0] exited with code [254] via signal [SIGINT]
PM2        | 2022-07-23T10:23:59: PM2 log: App [sasa_front:0] starting in -fork mode-
PM2        | 2022-07-23T10:23:59: PM2 log: App [sasa_front:0] online
PM2        | 2022-07-23T10:23:59: PM2 log: App [sasa_front:0] exited with code [254] via signal [SIGINT]
PM2        | 2022-07-23T10:23:59: PM2 log: App [sasa_front:0] starting in -fork mode-
PM2        | 2022-07-23T10:23:59: PM2 log: App [sasa_front:0] online
PM2        | 2022-07-23T10:23:59: PM2 log: App [sasa_front:0] exited with code [254] via signal [SIGINT]
PM2        | 2022-07-23T10:23:59: PM2 log: App [sasa_front:0] starting in -fork mode-
PM2        | 2022-07-23T10:23:59: PM2 log: App [sasa_front:0] online
PM2        | 2022-07-23T10:24:00: PM2 log: App [sasa_front:0] exited with code [254] via signal [SIGINT]
PM2        | 2022-07-23T10:24:00: PM2 log: App [sasa_front:0] starting in -fork mode-
PM2        | 2022-07-23T10:24:00: PM2 log: App [sasa_front:0] online
PM2        | 2022-07-23T10:24:00: PM2 log: App [sasa_front:0] exited with code [254] via signal [SIGINT]
PM2        | 2022-07-23T10:24:00: PM2 log: Script /usr/bin/npm had too many unstable restarts (16). Stopped. "errored"
PM2        | 2022-07-23T10:25:48: PM2 log: Stopping app:sasa_front id:0
PM2        | 2022-07-23T10:25:48: PM2 error: app=sasa_front id=0 does not have a pid
PM2        | 2022-07-23T10:25:53: PM2 log: App [sasa_front:0] starting in -fork mode-
PM2        | 2022-07-23T10:25:53: PM2 log: App [sasa_front:0] online

Start by reproducing the issue with the provided apps.json and the pm2 startup/save sequence, then inspect the startup restoration path that handles the configured cwd after a hard reboot. Done means the app is restored with /home/admin/apps/sasa_front as its cwd and no longer enters the unstable-restart loop.