Netflix/atlas ? reverse-engineered prompt

Build me a Scala backend for managing dimensional time series data, something like a small in memory metrics database that can ingest data points, tag them with dimensions, and let me query and aggregate them over time.

I want it to be runnable locally with a simple web API, and also have a standalone mode so I can try it out without a lot of setup. Include a few basic endpoints for writing data, reading it back, and searching or filtering by dimensions. If it makes sense, add some example configuration and a tiny demo dataset so I can see it working right away.

Please keep the project organized so the core time series logic is separate from the API layer, and make sure it’s easy to test. If you need to check the latest library docs online, go ahead and do that.