Proxmox 9.0.10 - Slow API call

[megachips]

Hi there,

I use the proxmox api to Provision VMs. I'm just evaluating 9.0.10 now.

When making a call to the https://<endpoint>/api2/json/nodes/<hypervisor name>/qemu it's taking upwards of 4 seconds to respond. This used to take 0.1s on 8.x

Api call taking over 4 seconds on 9.0.10 (This used to take 0.1s before I upgraded)

> time curl -s -H 'Authorization: <token>' --insecure https://172.16.16.104/api2/json/nodes/bigchug/qemu > /dev/null
curl -s -H  --insecure https://172.16.16.104/api2/json/nodes/bigchug/qemu >   0.01s user 0.01s system 0% cpu 4.272 total

Api call on other hypervisor much faster

> time curl -s -H 'Authorization:<token>' --insecure https://172.16.32.104/api2/json/nodes/gruntbox/qemu > /dev/null
curl -s -H  --insecure https://172.16.32.104/api2/json/nodes/gruntbox/qemu >   0.01s user 0.01s system 15% cpu 0.119 total

Here are the hypervisor versions:

> curl -s -H 'Authorization: <token>' --insecure https://172.16.16.104/api2/json/nodes/bigchug/version
{"data":{"repoid":"deb1ca707ec72a89","release":"9.0","version":"9.0.10"}}                                       
> curl -s -H 'Authorization: <token>' --insecure https://172.16.32.104/api2/json/nodes/gruntbox/version
{"data":{"repoid":"2a5fa54a8503f96d","version":"8.4.1","release":"8.4"}}

What could be the issue here? Why is it taking so long?

 

[Hannes Laimer]

Hey,

this endpoint hasn't changed in a while, actually quite a bit longer than the latest major release. Is there any chance something changed network-wise? And do you notice the same when doing

pvesh get /version --output-format json

on the node itself?

 

[megachips]

Hey,

this endpoint hasn't changed in a while, actually quite a bit longer than the latest major release. Is there any chance something changed network-wise? And do you notice the same when doing

pvesh get /version --output-format json

on the node itself?

Nothing has changed network wise. All I did was upgrade the node.

Here is a comparison using pvesh on an 8.4.1 node and 9.0.10 node.

 

[megachips]

Interesting - I hadn't rebooted the hypervisor since the upgrade.

After rebooting - it responds like it should