API disk, diskread, diskwrite report '0'

[Davis]

Hi, I am gathering stats from Proxmox 3.x and 4.x hosts using the JSON API (api2). However, all my VMs have 'diskread', 'diskwrite', and 'disk' reporting '0'.

I have confirmed it's not just my API client either, as running `pvesh` and doing a `get /nodes/{node}/qemu` also shows those properties as '0'.

When I do a `qm status {vmid} --verbose`, I can see the correct diskread and diskwrite values.

Is there a reason these would be reporting 0? I've looked through the existing Proxmox API documentation, and other forum posts, but have found nothing.

 

[dcsapak]

we only get those stats with a full status call
which is (for each single vm)
/nodes/<NODENAME>/qemu/<VMID>/status/current

or for all vms (but beware of the performance implications)
/nodes/<NODENAME/qemu
with parameter 'full 1'
so e.g. with pvesh

pvesh get /nodes/NODE/qemu --full 1

this is a costly operation since it connects to the qemu monitor several times for each vm

 

[Davis]

this is a costly operation since it connects to the qemu monitor several times for each vm

Thank you for the information. What is this 'costly' in terms of--the API, the VMs, only time...?

 

[dcsapak]

What is this 'costly' in terms of--the API, the VMs, only time...?

i mean it is time consuming(the process has to connect to every qemu process of every vm on that node, which may take a "while", compared to the rest of the api call), so it is not the default