[FEATURE]Service time granularity sampling storage

[YANG-DB]

Is your feature request related to a problem?

I would like to allow users to defined a pre-determining strategy for storing service based dimensions sampling (time-series data point)
This strategy will allow the user's understanding of the domain to define lower granularity storage strategy for services logs that are less critical or less prone to errors.

What solution would you like?

1. Allow the user to define for each service centric logs related monitoring storage indices - what time granularity to save the data-points into and which aggregation function to use for that (Avg, Max, Min, Sum)
2. Once defined, the target index that is generated for each service dimension measurement (CPU, Memory, UserCount ) would be using the Index Rollup job to auto aggregate and compact the original measurement index into.
3. The original measurement index info would be scraped using the Index State Management so that the user dosent have to manually delete the raw data.

Do you have any additional context?

• Index State Management
• OTEL serviceName
• Proposal: Supporting Real User Monitoring Events in OpenTelemetry open-telemetry/oteps#169