What do you use for "big" time-series data with a lot of tail reads?

Right now we have gotten away with using MS SQL Server but with the time-series data table reaching 30 TB and taking up 99.999...% of the database it might be time to look for an alternative.

A few years ago we ignored the purpose built time-series databases like InfluxDB and TimescaleDB and instead looked at the fully FOSS Apache Cassandra but we instead decided to squeeze a few years more out of the SQLServer database.

What are you using to store a lot of time-series data where the most recent month gets queried intensely and the rest gets queied a lot?