Pre-trained software containers for datacenter analysis

1. A method comprising:

discovering, by an onsite computing device, devices that are part of a datacenter;

creating, by the onsite computing device, a relationship amongst a group of devices from the discovered devices, based on a criterion;

downloading, by the onsite computing device, a pre-trained software container to each device of the group from a cloud system, wherein the pre-trained software container is specific to respective device the pre-trained software container is downloaded to, and

wherein the pre-trained software container is pre-trained to collect telemetry data from the respective device, identify an anomaly related to the respective device from the telemetry data, and generate an inference in a pre-defined format;

receiving, by the onsite computing device, the inference in the pre-defined format from respective pre-trained software container; and

analyzing, by the onsite computing device, the received inference.

2. The method of claim 1 , wherein the pre-trained software container collects the telemetry data from an agent running on the respective device.

3. The method of claim 1 , wherein the pre-trained software container parses the telemetry data into a format that enables identification of the anomaly related to the respective device.

4. The method of claim 1 , wherein the pre-trained software container includes an analytical model to identify the anomaly related to the respective device.

5. The method of claim 1 , wherein the criterion includes that the group of devices are associated with a specific tenant of the datacenter.

6. The method of claim 1 , wherein the criterion includes that the group of devices are related to a Service Level Agreement (SLA).

7. The method of claim 1 , wherein downloading includes downloading the pre-trained software container to a virtual machine (VM) on the respective device.

8. A system comprising:

a processor; and

a machine-readable medium storing instructions that, when executed by the processor, cause the processor to:

discover devices that are part of a datacenter;

create a relationship amongst a group of devices from the discovered devices, wherein the group of devices are associated with a specific tenant of the datacenter;

download a pre-trained software container to each device of the group from a cloud system, wherein the pre-trained software container is specific to respective device the pre-trained software container is downloaded to, and

wherein the pre-trained software container is pre-trained to collect telemetry data from the respective device, identify an anomaly related to the respective device from the telemetry data, and generate an inference in a pre-defined format;

receive the inference in the pre-defined format from respective pre-trained software container; and

analyze the received inference.

9. The system of claim 8 , wherein the pre-trained software container is pre-trained to interact with other pre-trained software containers on respective devices to identify an anomaly related to a device within the group.

10. The system of claim 8 , wherein the pre-trained software container shares information related to the anomaly with other pre-trained software containers in the datacenter.

11. The system of claim 8 , wherein the criterion includes that the group of devices are capable of generating telemetry data.

12. The system of claim 8 , wherein the pre-trained software container includes an analytical model to identify the anomaly related to the respective device.

13. The system of claim 12 , wherein the analytical model is developed within the datacenter based on data collected from the devices that are part of the datacenter.

14. The system of claim 8 , wherein the anomaly is identified through an analytical model trained within the datacenter, based on data collected from the devices that are part of the datacenter.

15. A non-transitory machine-readable storage medium comprising instructions, the instructions executable by a processor to:

discover devices that are part of a datacenter;

create a relationship amongst a group of devices from the discovered devices, based on a criterion;

download a pre-trained software container to each device of the group from a cloud system, wherein the pre-trained software container is specific to respective device the pre-trained software container is downloaded to, and

wherein the pre-trained software container is pre-trained to collect telemetry data from the respective device, identify an anomaly related to the respective device from the telemetry data, and generate an inference in a pre-defined format;

receive the inference in the pre-defined format from respective pre-trained software container; and

analyze the received inference.

16. The storage medium of claim 15 , wherein the analytical model is trained within the datacenter based on data collected from the devices that are part of the datacenter.

17. The storage medium of claim 15 , wherein the cloud system is one of a public cloud system, a private cloud system, and hybrid cloud system.

18. The storage medium of claim 15 , wherein the pre-trained software container is pre-trained to discover other pre-trained software containers in the datacenter.

19. The storage medium of claim 17 , wherein the pre-trained software container builds a relationship with at least one of the discovered pre-trained software containers.

20. The storage medium of claim 15 , wherein the criterion includes that the group of devices are associated with a specific tenant of the datacenter.