go.etcd.io/etcd@v3.3.27+incompatible/Documentation/benchmarks/etcd-3-demo-benchmarks.md (about) 1 --- 2 title: Benchmarking etcd v3 3 --- 4 5 ## Physical machines 6 7 GCE n1-highcpu-2 machine type 8 9 - 1x dedicated local SSD mounted under /var/lib/etcd 10 - 1x dedicated slow disk for the OS 11 - 1.8 GB memory 12 - 2x CPUs 13 - etcd version 2.2.0 14 15 ## etcd Cluster 16 17 1 etcd member running in v3 demo mode 18 19 ## Testing 20 21 Use [etcd v3 benchmark tool][etcd-v3-benchmark]. 22 23 ## Performance 24 25 ### reading one single key 26 27 | key size in bytes | number of clients | read QPS | 90th Percentile Latency (ms) | 28 |-------------------|-------------------|----------|---------------| 29 | 256 | 1 | 2716 | 0.4 | 30 | 256 | 64 | 16623 | 6.1 | 31 | 256 | 256 | 16622 | 21.7 | 32 33 The performance is nearly the same as the one with empty server handler. 34 35 ### reading one single key after putting 36 37 | key size in bytes | number of clients | read QPS | 90th Percentile Latency (ms) | 38 |-------------------|-------------------|----------|---------------| 39 | 256 | 1 | 2269 | 0.5 | 40 | 256 | 64 | 13582 | 8.6 | 41 | 256 | 256 | 13262 | 47.5 | 42 43 The performance with empty server handler is not affected by one put. So the 44 performance downgrade should be caused by storage package. 45 46 [etcd-v3-benchmark]: ../../tools/benchmark/