|
Post by dbaehr on Jul 7, 2017 15:17:01 GMT
Does the standard define any mechanism for querying performance data? I haven't come across anything in the docs yet but I wanted to be sure. I've seen perf data handled in many different ways across vendor APIs and it can be frustrating and time consuming to learn what options are available for filtering/generating this kind of information and documentation is usually poor, at best. This is typically a performance specific service endpoint that allows for filtering using resource properties (iops, cycles, bytes, temp, ...) and further filtering with specifiers like - granularity
- time frame options
- limit (number of values to return)
- aggregation functions (avg, min, max,...)
|
|
|
Post by mraineri on Jul 12, 2017 19:39:29 GMT
Performance metrics like that are typically not something readily available to the management controller on a system, so we have not focused time on how to model this. Typically something like this is best to be retrieved in and in-band manner. Is there a particular application that you're considering for this sort of modelling? For the pieces about how to capture things in terms of averages, over periods of time, etc, we do have a Metrics Model proposal available for review (which is something we've been considering for things like power and thermal measurements). You can find that proposal here: www.dmtf.org/sites/default/files/standards/documents/DSP-IS0002_0.9a.zip
|
|