[FEATURE] Extract common data models + OS Collectors + Service Metrics

[khushbr]

Background Context
Performance Analyzer (Writer) plugin captures the system behavior at a fine grain(5s) level in the form of metrics and stats, which is periodically flushed to a shared memory location. The RCA Agent(Reader) component scans this directory for updates, parses the raw data into sqlite entries and populates the on-disk sqlite metricDB.

Problem Statement & Proposed Solution
In the past, a conscious design choice to segregate the OpenSearch independent code into a separate process(RCA) was made. This allowed to deploy changes to RCA without restarting the critical OpenSearch process. However, the code split wasn't done cleanly. The shared data models of the various metrics are currently part of the RCA codebase. This proposal is to address this circular dependency between PA and RCA, extracting out the common data models from RCA codebase to a shared library performance-analyzer-commons, this can in-turn be consumed by the 2 components.

[khushbr]

Sister issue on RCA: opensearch-project/performance-analyzer-rca#307

[khushbr]

Release done for PA-Commons 1.0.0 - https://github.com/opensearch-project/performance-analyzer-commons/releases/tag/1.0.0