Performance Habits

Measure first

Most slow APIs are slow because of the database or a remote call, not C#. Fix N+1 queries, add indexes, use async everywhere, and only then consider caching. Guessing without measuring wastes time.

Output caching and compression

Cache a read-heavy, rarely changing endpoint for 30 seconds and compress responses. Never cache user-specific data under a shared key.

builder.Services.AddOutputCache();
builder.Services.AddResponseCompression();

var app = builder.Build();
app.UseResponseCompression();
app.UseOutputCache();

app.MapGet("/catalog", () => Catalog.LoadAll())
   .CacheOutput(p => p.Expire(TimeSpan.FromSeconds(30)));

Use health checks

Map app.MapHealthChecks("/healthz") so load balancers and Kubernetes can tell whether your app is alive. Add structured logging and metrics (OpenTelemetry) so you can see problems before users report them.

Quick check: What should you do before adding caching to a slow endpoint?

• Rewrite it in another language
• Turn off logging
• Add more middleware
• Measure to find the real bottleneck