Hi all,
just a small-fun question (?) maybe. I note in Proxmox webUI, since (a very long time now) we have easy access to "Disk IO" metrics, that are typically rendered as a 1-hour graph (RRD style) when we are in the 'summary' view of a VM. (please see small attached "metrics" screen grab to see what I mean more clearly?)
However, the graphical view .. is 1hr ... is a graphic object .. is somewhat limited as a result.
I was curious if anyone knows, is there an 'easy' way to harvest this data for a longer term ?
for example,
-- capture the IO numbers for all Vms on a single proxmox host, for a 24-hour period ?
-- store to a text (etc) file, which can then be opened and graphed separately / later ? or at least let me do meaningful comparisons without too much drama ?
My general goal here is to get a feel for how all the VMs on a few different proxmox hosts - are using the storage - in terms of
- who is a high-IO demand VM
- who is medium demand
- what the patterns look like
in theory I can just .. open all the summary pages once per hour, do some screenshots, and lots of copy-and-paste .. into a big old document ... and put things side by side .. and kludge it together. But that is kind of gross.
So I was just wondering if anyone knows this info is available, more directly somehow or not.
(ie, underlying KVM API / system level means? whatever is feeding the data to the thing that does the rendering for the proxmox GUI for the summary page ?)
Thanks!
Tim
just a small-fun question (?) maybe. I note in Proxmox webUI, since (a very long time now) we have easy access to "Disk IO" metrics, that are typically rendered as a 1-hour graph (RRD style) when we are in the 'summary' view of a VM. (please see small attached "metrics" screen grab to see what I mean more clearly?)
However, the graphical view .. is 1hr ... is a graphic object .. is somewhat limited as a result.
I was curious if anyone knows, is there an 'easy' way to harvest this data for a longer term ?
for example,
-- capture the IO numbers for all Vms on a single proxmox host, for a 24-hour period ?
-- store to a text (etc) file, which can then be opened and graphed separately / later ? or at least let me do meaningful comparisons without too much drama ?
My general goal here is to get a feel for how all the VMs on a few different proxmox hosts - are using the storage - in terms of
- who is a high-IO demand VM
- who is medium demand
- what the patterns look like
in theory I can just .. open all the summary pages once per hour, do some screenshots, and lots of copy-and-paste .. into a big old document ... and put things side by side .. and kludge it together. But that is kind of gross.
So I was just wondering if anyone knows this info is available, more directly somehow or not.
(ie, underlying KVM API / system level means? whatever is feeding the data to the thing that does the rendering for the proxmox GUI for the summary page ?)
Thanks!
Tim
