optimize-performance

To reduce backend latency, you must first measure the current endpoint speed, profile to pinpoint bottlenecks like N+1 queries, then design changes such as adding indexes or caching before verifying the improvement.

The best way to profile backend performance is to follow a measured workflow that identifies slow paths in endpoints, database queries, or microservices pipelines, pinpointing CPU, I/O, or N+1 issues rather than guessing.

No, you should not optimize microservices performance without measurement. The workflow enforces safety rules requiring you to collect metrics and identify the slow path before proposing and implementing any caching or indexing changes.

To apply database indexing, measure the current latency, profile the query bottleneck, then implement the index alongside batching or caching while preserving behavior, and verify the endpoint improvement using existing tests.

You choose between caching and indexing after profiling the bottleneck. Profiling determines whether latency stems from I/O, CPU, or database queries, guiding you to propose one or two concrete changes that best address the specific slow path.

Yes, backend performance optimization applies to microservices pipelines. It targets web services and database-heavy queries where latency matters, enforcing parameterized queries and pagination to safely reduce latency across the pipeline.