# API Latency Distribution

Twelve response-time samples per backend service show that `cache-service` is tight and fast while `search-service` carries a heavy long tail, exactly the spread a violin makes visible where a single average would hide it.