Separating a single database into one user/transaction database and one time series database

My project involves storing sensor data (time series data) in db. There are about 5,000 sensors, each is sending data at 15 min to 1 hour interval. The existing service is using a single db, storing all frontend user, transaction and sensor data. The table holding time series data now has over 10 million rows. Other tables such as user table has a few thousands rows. If separating sensor data into a dedicate db (say Postgres with TimescaleDB extension), will there be significant benefits in terms of performance, that out-weights the complexity introduced?