TimescaleDB 2.11.1 is installed as single node (single computer) on PostgreSQL 15.3 on Red Hat 8.8.
In my biggest hypertable I have five companies data to manage. Currently we have one single table with the following table fields (simplified sample):
Data from individual company is never mixed with other company data (those companies are strictly separated). In SQL statements there is always
WHERE company = 1 and ... condition (1 is for company 1, 2 is for company 2 etc).
Now I am thinking of somehow separate company data in separate categories.
- create PostgreSQL instance data separated by company (advantage is I can quickly create new Linux and then separate data into several virtual machines),
- create hypertable for each of the company,
- create space dimension (data are partitioned by time and “space”).
What is bothering me is third option. In documentation it is written that primary intention for “space” dimension is for multinode (multiple computers) configuration and it is not recommended for single node installations. Space dimensions on single node should only be used in special cases.
I am wondering if my case is such special case when space dimensioning would be beneficial? The benefit of “space” dimension would be no applications changes required.
Is it smart to do a space dimensioning in my case?