PostgreSQL vs ClickHouse: Filtering and Sorting
Introduction
Using a database-per-service approach in a microservice architecture is standard practice for good reason. It allows each microservice to be independently maintainable and deployable. The downside appears when a client needs a single datagrid assembled from multiple services, with sorting and filtering across any column.
In that model, one service has to fetch data from the others, assemble the combined result set in memory, then sort, filter, and page before returning a response. That can work at small scale, but it does not scale well as row counts grow.
