Big data query performance


At our company, we have a table with more than 1kkk rows on PostgreSQL, which we need to use for metrics development. Although Metabase may not be a suitable tool to perform such questions (I found the framework more suitable), I would like to ask your team some advice to work around such constraint.

Hi @brunolnetto

I don't know what "1kkk rows" is supposed to mean. 1 billion?
Metabase makes queries against your database, so it's up to your database to make the actual query.
I've seen Metabase connected to warehouses with many millions of rows.

You'll need to be a lot more specific about your performance problem.

The architecture here it given as follows:

Entity A represents the top-hierarchy level: from it, we represent children entity B binded to entity A, in a 1:n relation. Subsequently, entity B exhibits state steps change entity C, like a state machine, also in a 1:n relation. Each step entity, A, B and C, are recorded in a relational PostgreSQL database, which makes the relation A-C somewhat sum_ij 1:n_i*n_j, for index i respective to entity B and index j respective to entity C.

The aforementioned table refers to entity C table. In our case, entity B table has already 40 million rows, which gives a dimensionality sense of the problem.

@brunolnetto Okay, that sounds like database problem, not a Metabase problem.

Would you provide some advice of handling such issue?

@brunolnetto You should contact a DBA to get advise about database.