Hierarchical tree of time series with different update rates

Hi, I’m trying to find some information to guide me on the best way to organize my data for long term archiving/storage.

We have a Redis based data exchange set up that uses a materialized tree approach to implement exchange of hierarchical data between different system and antennas of a millimeter-wave interferometric telescope. This works well for exchanging current values (and using pub/sub notifications to trigger certain actions). The hierarchy of data can be moderately deep - say 10-15 levels in places, and values for specific data items can change at rates between a bit less than a second to days/years.

(Background info: historically we’ve had a flat table of data and an SQL database has worked OK with a home built perl based plotting system. We’re now moving to this more modular/hierarchical set up to try and make things easier to maintain as the system gets more complicated)

What we need to do now is log and archive that data for many years in a way that is fairly easy query, e.g. to explore problems with the telescope operations, or to understand the state of the instrument during a historical observation. MongoDB time series, linked to Grafana or similar seem like a reasonably good fit for this application.

What is the best way to structure hierarchical trees of time series with varying cadences? A materialized tree of time series collections? Or a time series of materialized trees? Or some hybrid between the two? Long term storage efficiency and querying time series of a collection of specific values in the tree are probably the two main driving use cases/scenarios.