[FEATURE-REQUEST] Proxmox VE scheduled tasks

[xtonousou]

It would be very useful for sysadmins and devops teams, if there was an embedded task scheduler/queue worker for native actions like QEMU/LXC power actions, migrations and deletions with system and user defined intervals or just once-off actions. This feature should be available through API and compatible with standalone or cluster deployments.

 

[cheiss]

Hi,

for feature requests, bug reports etc. we have a dedicated tracker over at https://bugzilla.proxmox.com/!
You can create a new ticket there and lay out your use case in detail. This helps us, as well as yourself, to properly keep track of the status. Otherwise it can get lost here in the forum pretty easily.

 

[LnxBil]

It would be very useful for sysadmins and devops teams, if there was an embedded task scheduler/queue worker for native actions like QEMU/LXC power actions, migrations and deletions with system and user defined intervals or just once-off actions. This feature should be available through API and compatible with standalone or cluster deployments.

Those jobs could not be triggered via SSH or simple cronjobs? If not, what are those jobs, can you given an example?

 

[xtonousou]

Those jobs could not be triggered via SSH or simple cronjobs? If not, what are those jobs, can you given an example?

For example (The most important ones are the scheduled tasks for Virtual Machines and LXCs):

• Virtual Machine or LXC
  
  • Power On/Off
  • Shut Down Guest OS
  • Restart Guest OS
  • Suspend
  • Reset/Reboot
  • Take snapshot
  • Migrate
  • Delete
    • with "Purge from job configurations"
    • with "Destroy unreferenced disks owned by guest"
  • Clone (maybe?)
• PVE Hosts:
  • New Virtual Machine or LXC
• PVE Cluster:
  • New Virtual Machine or LXC (based on available resources)
  • Add PVE Host

This functionality is implemented on VMware vCenter product (for comparison):


As you mentioned, these jobs could be triggered via external factors (e.g. Ansible, Cronicle, Semaphore, Kestra, n8n), but I think the native/embedded functionality would be more sustainable and useful for the admins and for the PVE itself. I understand the fact that PVE does not provide an external management instance like vCenter, but I think that the implementation of scheduled tasks is possible by just utilizing Systemd and the clustered filesystem or with the same way that backup scheduling is implemented.

 

[LnxBil]

Thank you for the explanation.