[floppydisk]

I managed >1TB Postgres installs using not-fancy tools and we did just fine. The How:

Installing databases was pretty easy. We wrote DDLs for our databases and created shell scripts that would leverage pgsql via the command line to create the databases with all our desired extensions, plugins, functions, and schemas. Repeatable every time by even the least trained IT guy on the staff. Data loading was a little more finicky, but that was easily doable by using pgdump if we were coming from an existing database and it was executable by shell script.

Stats wise, we wrote stored functions in the database that would leverage the built in, and excellent, stats API. Our management system would execute a shell script that leveraged PGSQL to execute the stored procedure(s) and pass the resulting values to our graphing and system monitoring software. Standard setup on each DB box. Worked like a champ.

For managing configuration files, we preferred to be hands on and edit pgconf directly as each database box ended up being a little different in terms of needs and we would annotate the configuration file with notes to selves about why settings were the way they were.

How do other services do this via console? They interface with the database API directly (like you can do) and make an interface to trigger the commands they're executing.