How do I fix slow aggregation, pagination or N+1 queries?

Definition:

Query performance depends on how much data is scanned, joined, sorted and returned for each useful result. Slow pagination and N+1 patterns often do repeated work that grows with result size. The best fix usually reduces that work before increasing hardware or connection limits.

Key points:

• Check scan counts and indexes; reduce documents before expensive joins or groups.

• For deep pages, use a stable cursor such as createdAt plus _id.

• Batch related reads and reuse a bounded connection pool.

• Keep filters, selected fields and sort keys consistent with cursor construction, including a unique tie-breaker.

Interview point:

• Compare query count, examined documents and latency before and after the change, not just endpoint code length.