We used standard Postgres for 200k rows/day (100M rows/year) time series. Worked just fine, even without partitioning. Maybe use partitioning (on time) from the

We used standard Postgres for 200k rows/day (100M rows/year) time series. Worked just fine, even without partitioning. Maybe use partitioning (on time) from the start though, it was a bit tedious to migrate to it later. Now handling 10x these rates, probably good for another 10-100x. We have TimeScaleDB extension enabled now, but not using that many of the features yet, mostly stock Postgres.