issue:#4043165: Create a new telemetry_collector class

[egershonNvidia]

What

Separate the telemetry retrieval code into its own class in a different file.
Add a class for storing information, which saves the telemetry data and its deltas.

Why ?

Task
The second task, in order to merge the ALM and PDR plugins, is to move the telemetry to its class, decoupling the main algorithm components to prepare for the merge with ALM.

How?

took get_telemetry from ufm communication manager, put it in its class, and added a delta function to extract the delta between 2 telemetries. in another class I have put all the data store-related functions: save, remove old, and locations of folders of saved files, depending on absolute or delta. This was taken from ALM but I have put 10 items to keep, not for a particular reason. PDR plugin does not use those files, only for debugging.

Every main loop iteration, the server checks the data store to make sure that we don't overflow the data, with clean_old_files.

Testing

regular testing needs to run. Added a test to see that those files were created in the latest 5 minutes.
Added a test to make sure the data store is cleaned up and only maximum items are there.

Special triggers

Use the following phrases as comments to trigger different runs

• bot:retest rerun Jenkins CI (to rerun GitHub CI, use "Checks" tab on PR page and rerun all jobs)
• bot:upgrade run additional update tests

[TalLerner]

Can you please rephrase the text you wrote in the why section of the PR's description?
You can explain the motivation for decoupling the main algorithm parts to prepare it for the merge with ALM.
By the way, the removal of the dynamic telemetry is a better implementation regardless of the merge with the ALM. The dynamic telemetry mechanism wasn't the best choice for the PDR.