Technologies for capturing processing resource metrics as a function of time
Technologies for collecting metrics associated with a processing resource (e.g., central processing unit (CPU) resources, accelerator device resources, and the like) over a time window are disclosed. According to an embodiment presented herein, a network device receives, in an edge network, a request to provide one or more metrics associated with a processing resource, the request specifying a window indicative of a time period to capture the one or more metrics. The network device obtains the one or more metrics from the processing resource for the specified window and provides the obtained one or more metrics in response to the request.
1. A network device comprising:
memory; and
circuitry to:
receive, in an edge network, a request to provide one or more metrics associated with a processing resource, the request specifying a window indicative of a time period to capture the one or more metrics;
obtain the one or more metrics from the processing resource for the specified window, wherein to obtain the one or more metrics from the processing resource includes to:
determine, from the request, a usage context indicative of a use for the obtained one or more metrics;
identify, as a function of the usage context, a subset of the one or more metrics; and
cause the processing resource to activate one or more performance counters associated with each metric in the identified subset; and
collect the one or more metrics from each of the activated performance counters over the specified window; and
provide the obtained one or more metrics in response to the request.
2. The network device of claim 1 , wherein the circuitry is to receive the request to provide the one or more metrics associated with a plurality of processing cores, the request specifying the window indicative of the time period to capture the one or more metrics.
3. The network device of claim 1 , wherein the circuitry is to receive the request to provide the one or more metrics associated with a plurality of accelerator devices, the request specifying the window indicative of the time period to capture the one or more metrics.
4. The network device of claim 1 , wherein the circuitry is to provide the obtained one or more metrics in response to the request by storing the obtained one or more metrics in a repository for retrieval by one or more machine learning elements.
5. The network device of claim 1 , wherein the circuitry is to provide the obtained one or more metrics in response to the request by transmitting the obtained one or more metrics in response to the request.
6. The network device of claim 1 , wherein the circuitry is to identify, as a function of the usage context, the subset of the one or more metrics by applying a filtering scheme to the one or more metrics specified in the request based on the usage context.
7. The network device of claim 1 , wherein to obtain the one or more metrics from the processing resource further includes to deactivate each of the performance counters in response to one or more events.
8. A method comprising:
receiving, in an edge network, a request to provide one or more metrics associated with a processing resource, the request specifying a window indicative of a time period to capture the one or more metrics;
obtaining the one or more metrics from the processing resource for the specified window by:
determining, from the request, a usage context indicative of a use for the obtained one or more metrics;
identifying, based on the usage context, a subset corresponding to the one or more metrics;
causing the processing resource to activate one or more performance counters associated with each metric in the subset; and
collecting the one or more metrics from each of the activated performance counters over the specified window; and
providing the one or more metrics in response to the request.
9. The method of claim 8 , wherein the one or more metrics of the request are associated with a plurality of processing cores.
10. The method of claim 8 , wherein the one or more metrics of the request are associated with a plurality of accelerator devices.
11. The method of claim 8 , wherein the identifying of the subset corresponding to the one or more metrics is based on a filtering scheme corresponding to the usage context.
12. The method of claim 8 , wherein the obtaining of the one or more metrics from the processing resource includes deactivating each of the performance counters in response to one or more events.
13. The method of claim 8 , further including providing the obtained one or more metrics in response to the request by storing the obtained one or more metrics in a repository for retrieval by one or more machine learning elements.
14. One or more non-transitory machine-readable storage media storing instructions, which, when executed on one or more processors, cause a network device to at least:
receive, in an edge network, a request to provide one or more metrics associated with a processing resource, the request specifying a window indicative of a time period to capture the one or more metrics;
obtain the one or more metrics from the processing resource for the specified window by:
determining, from the request, a usage context indicative of a use for the obtained one or more metrics;
identifying, as a function of the usage context, a subset of the one or more metrics;
causing the processing resource to activate one or more performance counters associated with each metric in the identified subset; and
collecting the one or more metrics from each of the activated performance counters over the specified window; and
provide the obtained one or more metrics in response to the request.
15. The one or more non-transitory machine-readable storage media of claim 14 , wherein the instructions are to cause the network device to receive the request to provide the one or more metrics associated with a plurality of accelerator devices, the request specifying the window indicative of the time period to capture the one or more metrics.
16. The one or more non-transitory machine-readable storage media of claim 14 , wherein to identify, as the function of the usage context, the subset of the one or more metrics, the instructions are to cause the network device to apply a filtering scheme to the one or more metrics specified in the request based on the usage context.
17. The one or more non-transitory machine-readable storage media of claim 14 , wherein to obtain the one or more metrics from the processing resource, the instructions are to cause the network device to
deactivate each of the performance counters.
18. The one or more non-transitory machine-readable storage media of claim 14 , wherein to provide the obtained one or more metrics in response to the request, the instructions are to cause the network device to store the obtained one or more metrics in a repository for retrieval.
19. The one or more non-transitory machine-readable storage media of claim 14 , wherein to provide the obtained one or more metrics in response to the request, the instructions are to cause the network device to transmit the obtained one or more metrics in response to the request.
20. The one or more non-transitory machine-readable storage media of claim 14 , wherein the instructions are to cause the network device to receive the request to provide the one or more metrics associated with a plurality of processing cores, the request specifying the window indicative of the time period to capture the one or more metrics.