bug: deleted vm telemetry pollutes new VM history

[genewitch]

https://i.imgur.com/18PYzoT.png
if you make vm 103, delete it, and then later make another VM that gets numberd 103, the new VM will show the old VM's graphs/telemetry
see attached. I don't remember what VM was 103 in the beginning of 2024, but it was decommissioned in september.
A new service was spun, i let proxmox just autoassign the ID number.

I can imagine a circumstance where this could be a liability, but really, it's just poor compartmentalization. I'm sure these VMs and CTs have UUIDs, so why would metrics/telemtery be saved under the equivalent of an 8 bit int?

 

[genewitch]

Note: this has nothing to do with the number 103, and everything to do with the VM ID number being used as the key.

 

[Kingneutron]

It's just the PVE GUI display, AFAIK nobody relies on it for actual metrics. It's just what-the-vm-is-doing performance at a glance; should scroll off in 24H or so, maybe less. Might be as little as an hour.

If you want real data to rely on you probably would implement e.g. Prometheus, influxdb, grafana, sar in-vm, etc.

https://search.brave.com/search?q=p...summary=1&conversation=599180280d158e976a89f7

But yes, if a VM is deleted they should probably cleanup the old data.

 

[genewitch]

It's just the PVE GUI display, AFAIK nobody relies on it for actual metrics. It's just what-the-vm-is-doing performance at a glance; should scroll off in 24H or so, maybe less. Might be as little as an hour.

If you want real data to rely on you probably would implement e.g. Prometheus, influxdb, grafana, sar in-vm, etc.

https://search.brave.com/search?q=p...summary=1&conversation=599180280d158e976a89f7

But yes, if a VM is deleted they should probably cleanup the old data.

thanks for considering to reply; did you click the image? that's the "yearly" graphs. Saying "you're holding it wrong" doesn't make it any less of a bug.