The Throughput shows how many requests your service is processing over time. Throughput is measured in RPCs per second (RPC/s), where each RPC (Remote Procedure Call) represents a single request–response cycle between the client and server. The graph helps you understand traffic patterns, spot sudden spikes or drops, and track how efficiently your system is handling load.
A steady throughput generally indicates stable performance, while fluctuations may point to scaling issues, bottlenecks, or unexpected traffic changes. Monitoring throughput alongside latency and error rate graphs gives a clearer picture of system health and helps in identifying performance regressions or overload conditions.