Why was reporting built on PostgreSQL under a load of ~2000 RPS? Why not use OLAP instead of OLTP for reporting?

Using PostgreSQL for reporting at around 2000 RPS may be due to several reasons:

• Infrastructure unification: Using a single type of database simplifies support, monitoring, and team training.
• Sufficient performance: PostgreSQL scales well and can efficiently handle high loads with proper configuration and indexing.
• Complexity and cost of OLAP: Implementing a specialized OLAP solution requires additional resources for development, support, and integration.
• Data requirements: If reports are built on fresh data with minimal delay, an OLTP system may be preferable, as OLAP often works with aggregated or historical data.

OLAP solutions are typically used for complex analytical analysis with large volumes of data and are less sensitive to response time. If reporting requires real-time data and high availability, OLTP on PostgreSQL might be a more suitable choice.

Thus, choosing PostgreSQL for reporting at 2000 RPS is a balance between performance, cost, and data requirements.