Databases for Backend Performances
Database performance work means making queries return quickly and keeping the database healthy as load rises. It includes reading query plans, designing indexes, avoiding N+1 patterns, pooling connections and scaling with replicas or partitioning. The database is the most common bottleneck in a backend, so this unit has the highest payoff.
Stage: Data Layer · Part of the Backend Performance roadmap · Last reviewed 19 November 2026
What to learn in Databases
- Read an EXPLAIN plan and spot a sequential scan on a large table
- Design a composite index that matches a query's filter and sort order
- Detect and remove an N+1 query pattern in application code
- Size a connection pool and explain what happens when it is exhausted
- Route read traffic to replicas while handling replication lag
Databases interview questions
Why can adding an index make writes slower?
Each index is a separate structure that must be updated on every insert, update or delete that touches its columns. More indexes mean more write work and more storage. Index the columns that real queries filter or sort on, and drop unused ones.
What is connection pooling and why is it needed?
A pool keeps a small set of open database connections that requests borrow and return. Opening a connection is costly, and databases cope badly with thousands of simultaneous ones. Pooling cuts connection setup time and caps load on the database.
Where Databases fits in the Backend Performance roadmap
Databases belongs to the Data Layer stage (advanced level), which takes about 14 hours. Usually the bottleneck.
Other units in this stage
Back to the full Backend Performance roadmap