Hello!
I have a database with circa 16B documents (and growing), with data from sensors. Right now, the data is divided in monthly non-time-series collections. We divided the documents this way to reduce de query delay. And if we need to query data from multiple months, we make multiple queries.
Now we are migrating to a time-series database, because it’s perfect for our use case, and can improve the performance.
So my question is, if it’s worth to still separate the data in monthly collections, or it’s better to keep all documents in a single collection, since the queries for data with the same metaField is optimized.