feat: Monitoring for Feast Feature Store (python server)

[tsisodia10]

What this PR does / why we need it:

This draft pull request introduces monitoring functionality for the Feast feature store(python server). It instruments the feature store to capture system-level metrics, currently focusing on CPU and memory usage. Additionally, this PR includes the setup for a Prometheus server to collect and store these metrics.

Note: This pull request is currently a work in progress.

[shuchu]

Is it possible to use open-telemetry client for instrumentation?

[jeremyary]

@shuchu we'll definitely do some investigating! We're not necessarily opposed to proper abstraction that could bridge components or something similarly if need be down the road. I saw that it's the second most-active CNCF project so I'm guessing it's fairly popular :)

[shuchu]

@shuchu we'll definitely do some investigating! We're not necessarily opposed to proper abstraction that could bridge components or something similarly if need be down the road. I saw that it's the second most-active CNCF project so I'm guessing it's fairly popular :)

Thank you Jeremy. The Open-telemetry seems to be a "standard" service for many companies. I thought it not only can benefit the Feast but also the Red Hat's OpenShift (https://www.redhat.com/en/technologies/cloud-computing/openshift/observability).

[tsisodia10]

Thanks @shuchu for the recommendation, that sounds like a great idea and we are looking into the implementation.

[shuchu]

Thanks @shuchu for the recommendation, that sounds like a great idea and we are looking into the implementation.

Thank you @tsisodia10 !