AIQUM volume latency data vs API data

[autumn timber]

Hi All,
from the image attached, fg_test has a spike latency with a value around 150ms on July 3, 2024 4:30PM,
The image is from AIQUM GUI
And here is the API output json from AIQUM API "/datacenter/storage/volumes/{key}/metrics" with interval 1w

The latency doesn't match the value in GUI, from the JSON,
"timestamp": "2024-07-03T17:00:00.000+08:00", latency value: 5635.39
"timestamp": "2024-07-03T16:00:00.000+08:00" , latency value: 5453.07

Also, is this value the maximum within an hour? Or this is the average value among one hour?

[gusty axle]

I wonder if that is a raw value or something? Like you have to divide by 300 or divide by IOPS?

I feel like this is something a case could answer.

Unless @sick ether is around?

[untold grove]

it should be raw, so maybe usec instead of ms?
for a 5 minute interval, i believe we take the highest of each 1 minute sample we receive from the cluster in that 5 minutes. for longer time periods like a week, i don't know if it's an average of the 5 minute samples we have, or if it should be taking the highest.

a case would be best i think.

[fiery nova]

I was just doing this kind of stuff today. I had to take a sample on AIQUM Rest (point in time) and then compare to what the WebUI was saying. I then had to divide by 1024/1024 a few times to get the right amounts; Eventually Sample A = Sample A .