Hello! We plan to run Greptime on an edge deployment which we own, so in practice it will be on premise. In the docs the preference goes to using managed SQL as a meta KV store. For us this is not an option. All advantages listed in the documenation do not apply for self managed Postgres. Is the SQL db still preferred if there is no managed solution available? Or is etcd a better option in this case?

We generally recommend using an SQL database as the backend for Metasrv, since it is much easier to maintain and offers a more robust disaster recovery solution. That said, this does not mean we don’t support etcd. The best choice ultimately depends on your team’s skills and expertise.

We are running a 3 node cluster so clustered has our preference. Ok good to know. We are already running postgres so we will use that!