Processes randomly error out upon reboot with kill_timeout defined

As described in the title, when kill_timeout is defined, there's a random chance PM2 will fail to restart the process, failing with the message [PM2][ERROR] Process 1 not found. My environment runs a python script that constantly runs, and reboots once the process reaches a memory limit monitored by PM2. I've defined kill_timeout because PM2's logs become spammy during reboots while it's waiting for the process to exit, before outright killing the process after a number of milliseconds. Since I'd prefer the process die immediately, I set kill_timeout to 80. This worked initially, but now this edge-case has become so prevalent that it's not possible to use this feature in a production environment as it leave at least one of my processes dead when restarting it automatically. The odd thing is that it's easier to reproduce this issue manually with the pm2 restart [app_name] command than it is with pm2 stop [app_name] and pm2 start [app_name]. Not certain what the issue is, though it seems like it may be timing related.

My pm2.yml is as follows:

apps:
  - name: 'app_name'
    cwd: '/path/to/app'
    script: '/path/to/app/.venv/bin/python'
    args: '...'
    watch: false
    instances: 1
    max_memory_restart: '10240M'
    wait_ready: true
    listen_timeout: 8000
    kill_timeout: 80
...

Let me know if further information is required. Thank you.

Start by reproducing the failure with the supplied pm2.yml, especially kill_timeout: 80, using pm2 restart [app_name] and comparing it with pm2 stop [app_name] followed by pm2 start [app_name]. Trace the restart behavior around the reported "Process 1 not found" error. Done means the configured process reliably restarts after reboot or restart without being left dead.