Terminology Guide
Key concepts, parameters, and indicators used in evaluating server and database performance.
Throughput
RPS (Requests Per Second)Measures the absolute number of successful requests resolved per second under load. Think of this as how many requests a server can process in one second.
Impact: Determines maximum concurrent users supported.Latency
Median Response (ms)The duration in milliseconds for a single request to travel to the server and back. Lower latency ensures instant UI feedback and highly snappy user interactions.
Impact: Vital for real-time mobile and web responsiveness.Tail Latency
p95 / p99 PercentileThe worst-case latency representing the slowest 5% of requests. Tells you if database queries, network jitter, or garbage collection are causing frustrating intermittent delays under load.
Impact: Guards against server-side bottlenecks and jitter.Peak Memory
RAM Footprint (MiB)The absolute maximum RAM overhead recorded during testing. Lower footprint yields cheap container orchestration (e.g. AWS ECS/Fargate) and seamless resource utilization.
Impact: Crucial for hosting cost-effectiveness and scaling.