|
|
Post by jmaaks on May 3, 2023 15:21:10 GMT
Hi all,
I'm trying to determine whether the Redfish API provides the capability for CPU and memory utilization reporting across a specific time period, and I'm not finding clear examples in the mockups, mockup server, or API documentation. It looks like TelemetryService might be the way to go, and the Redfish Telemetry White Paper hints that I'm in the ballpark.
Any pointers or feedback on this use-case for the Redfish API would be greatly appreciated!
|
|
|
|
Post by mraineri on May 3, 2023 16:05:33 GMT
You're absolutely going down the right path. The typical "metrics" resources (like MemoryMetrics and ProcessorMetrics for your utilization reporting) reflect the current state and don't necessarily have info on a per time window basis (other than perhaps a lifetime value). The telemetry service is intended to augment that to allow for reporting over time intervals, and automatic transmission of reports to clients that do not want to manually perform GET operations to fetch the reports.
It might be best to start with looking for your utilization metrics under ProcessorMetrics and MemoryMetrics as a baseline to understand the data coming back for the "current" values (specifically the BandwidthPercent properties), and then building on that to map it over to the telemetry service usage. We do have an example of "BandwidthPercent" in our public-rackmount1 mockup (under /redfish/v1/Systems/437XR1138R2/Processors/FPGA1/ProcessorMetrics); it doesn't look like we have any examples for memory though, but the property definition is very similar between the two.
|
|
