本文提供4核16G SSD 200G 3个节点的腾讯云 Elasticsearch Service 集群的性能指标。
说明：
对比4核16GB SSD 200G 3节点的腾讯云 ES 和社区版本的集群性能指标，腾讯云 ES 各方面性能均有一定优势。主要得益于腾讯云 ES 在底层存储模型优化（时序合并、冷分片持续合并）、查询执行计划优化（高效剪枝、缓存优化）、内置场景模板、自研 JDK、GC 参数调优等方面的优化，详见 内核增强。
geonames/7.10.1/4核16GB
说明 | Metric | | Unit | Task | 腾讯云 ES | 社区版本 | 差异值（社区版-ES） | 差异百分比(差异值/社区版) | 效果比较 |
写入总耗时 | Cumulative indexing time of primary shards | | min | - | 16.3633 | 17.8173 | 1.454 | 8.16% | 优 |
GC 总次数、耗时统计 | Total Young Gen GC time | | s | - | 6.26 | 68.454 | 62.194 | 90.86% | 劣 |
| Total Young Gen GC count | | 次 | - | 892 | 4163 | 3271 | 78.57% | 优 |
| Total Old Gen GC time | | s | - | 0 | 0 | 0 | - | - |
| Total Old Gen GC count | | 次 | - | 0 | 0 | 0 | - | - |
存储大小 | Store size | | GB | - | 2.51866 | 2.93883 | 0.42017 | 14.30% | 优 |
堆内存使用量 | Heap used for segments | | MB | - | 0.803783 | 0.709301 | -0.094482 | -13.32% | 劣 |
| Heap used for doc values | | MB | - | 0.0284767 | 0.0218697 | -0.006607 | -30.21% | 劣 |
| Heap used for terms | | MB | - | 0.655075 | 0.562836 | -0.092239 | -16.39% | 劣 |
| Heap used for norms | | MB | - | 0.0732422 | 0.0765991 | 0.0033569 | 4.38% | 优 |
| Heap used for points | | MB | - | 0 | 0 | 0 | - | - |
| Heap used for stored fields | | MB | - | 0.0469894 | 0.0479965 | 0.0010071 | 2.10% | 优 |
segment 总数量 | Segment count | | 次 | - | 6 | 97 | 91 | 93.81% | 优 |
写入吞吐、耗时统计 | Min Throughput | | docs/s | index-append | 89331.9 | 80608.3 | -8723.6 | -10.82% | 优 |
| Median Throughput | | docs/s | index-append | 90268.8 | 82576.8 | -7692 | -9.31% | 优 |
| Max Throughput | | docs/s | index-append | 90516.1 | 84030.5 | -6485.6 | -7.72% | 优 |
| 50th percentile latency | | ms | index-append | 233.258 | 305.781 | 72.523 | 23.72% | 优 |
| 90th percentile latency | | ms | index-append | 314.558 | 354.166 | 39.608 | 11.18% | 优 |
| 99th percentile latency | | ms | index-append | 341.303 | 403.776 | 62.473 | 15.47% | 优 |
| 100th percentile latency | | ms | index-append | 354.657 | 428.363 | 73.706 | 17.21% | 优 |
| 50th percentile service time | | ms | index-append | 233.258 | 305.781 | 72.523 | 23.72% | 优 |
| 90th percentile service time | | ms | index-append | 314.558 | 354.166 | 39.608 | 11.18% | 优 |
| 99th percentile service time | | ms | index-append | 341.303 | 403.776 | 62.473 | 15.47% | 优 |
| 100th percentile service time | | ms | index-append | 354.657 | 428.363 | 73.706 | 17.21% | 优 |
| error rate | | % | index-append | 0 | 0 | 0 | - | - |
index 指标统计 | Min Throughput | | ops/s | index-stats | 90.04 | 90.05 | 0.01 | 0.01% | 优 |
| Median Throughput | | ops/s | index-stats | 90.07 | 90.07 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | index-stats | 90.14 | 90.13 | -0.01 | -0.01% | 优 |
| 50th percentile latency | | ms | index-stats | 2.91003 | 3.07875 | 0.16872 | 5.48% | 优 |
| 90th percentile latency | | ms | index-stats | 3.82882 | 4.27689 | 0.44807 | 10.48% | 优 |
| 99th percentile latency | | ms | index-stats | 4.2378 | 4.75335 | 0.51555 | 10.85% | 优 |
| 99.9th percentile latency | | ms | index-stats | 4.34459 | 9.2331 | 4.88851 | 52.95% | 优 |
| 100th percentile latency | | ms | index-stats | 8.22393 | 17.3644 | 9.14047 | 52.64% | 优 |
| 50th percentile service time | | ms | index-stats | 1.78268 | 2.17806 | 0.39538 | 18.15% | 优 |
| 90th percentile service time | | ms | index-stats | 2.07484 | 2.5508 | 0.47596 | 18.66% | 优 |
| 99th percentile service time | | ms | index-stats | 2.43121 | 2.94067 | 0.50946 | 17.32% | 优 |
| 99.9th percentile service time | | ms | index-stats | 3.09198 | 3.50368 | 0.4117 | 11.75% | 优 |
| 100th percentile service time | | ms | index-stats | 7.29974 | 15.5761 | 8.27636 | 53.13% | 优 |
| error rate | | % | index-stats | 0 | 0 | 0 | - | - |
node 指标统计 | Min Throughput | | ops/s | node-stats | 90.06 | 90.05 | -0.01 | -0.01% | 优 |
| Median Throughput | | ops/s | node-stats | 90.09 | 90.09 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | node-stats | 90.34 | 90.33 | -0.01 | -0.01% | 优 |
| 50th percentile latency | | ms | node-stats | 3.17223 | 3.60429 | 0.43206 | 11.99% | 优 |
| 90th percentile latency | | ms | node-stats | 3.70681 | 4.1719 | 0.46509 | 11.15% | 优 |
| 99th percentile latency | | ms | node-stats | 5.01334 | 5.7717 | 0.75836 | 13.14% | 优 |
| 99.9th percentile latency | | ms | node-stats | 6.75018 | 7.29828 | 0.5481 | 7.51% | 优 |
| 100th percentile latency | | ms | node-stats | 7.98905 | 8.64352 | 0.65447 | 7.57% | 优 |
| 50th percentile service time | | ms | node-stats | 2.43876 | 2.80255 | 0.36379 | 12.98% | 优 |
| 90th percentile service time | | ms | node-stats | 2.78272 | 3.25325 | 0.47053 | 14.46% | 优 |
| 99th percentile service time | | ms | node-stats | 4.12234 | 5.21353 | 1.09119 | 20.93% | 优 |
| 99.9th percentile service time | | ms | node-stats | 6.35902 | 6.69629 | 0.33727 | 5.04% | 优 |
| 100th percentile service time | | ms | node-stats | 7.4313 | 7.52479 | 0.09349 | 1.24% | 优 |
| error rate | | % | node-stats | 0 | 0 | 0 | - | - |
默认查询，所有文档 score 为1（match_all） | Min Throughput | | ops/s | default | 50.03 | 50.03 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | default | 50.04 | 50.04 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | default | 50.08 | 50.07 | -0.01 | -0.02% | 优 |
| 50th percentile latency | | ms | default | 3.89929 | 4.97851 | 1.07922 | 21.68% | 优 |
| 90th percentile latency | | ms | default | 4.39236 | 5.47875 | 1.08639 | 19.83% | 优 |
| 99th percentile latency | | ms | default | 4.78834 | 6.31232 | 1.52398 | 24.14% | 优 |
| 99.9th percentile latency | | ms | default | 7.10486 | 46.1096 | 39.00474 | 84.59% | 优 |
| 100th percentile latency | | ms | default | 8.75822 | 59.3054 | 50.54718 | 85.23% | 优 |
| 50th percentile service time | | ms | default | 3.18269 | 4.13119 | 0.9485 | 22.96% | 优 |
| 90th percentile service time | | ms | default | 3.49347 | 4.47708 | 0.98361 | 21.97% | 优 |
| 99th percentile service time | | ms | default | 3.8746 | 5.30558 | 1.43098 | 26.97% | 优 |
| 99.9th percentile service time | | ms | default | 6.68581 | 9.82652 | 3.14071 | 31.96% | 优 |
| 100th percentile service time | | ms | default | 8.30396 | 58.1671 | 49.86314 | 85.72% | 优 |
| error rate | | % | default | 0 | 0 | 0 | - | - |
term 条件查询 | Min Throughput | | ops/s | term | 100.05 | 100.04 | -0.01 | -0.01% | 优 |
| Median Throughput | | ops/s | term | 100.07 | 100.06 | -0.01 | -0.01% | 优 |
| Max Throughput | | ops/s | term | 100.14 | 100.12 | -0.02 | -0.02% | 优 |
| 50th percentile latency | | ms | term | 3.17419 | 3.37056 | 0.19637 | 5.83% | 优 |
| 90th percentile latency | | ms | term | 3.62229 | 3.8708 | 0.24851 | 6.42% | 优 |
| 99th percentile latency | | ms | term | 4.03812 | 5.63819 | 1.60007 | 28.38% | 优 |
| 99.9th percentile latency | | ms | term | 5.9753 | 7.94086 | 1.96556 | 24.75% | 优 |
| 100th percentile latency | | ms | term | 8.03321 | 11.0892 | 3.05599 | 27.56% | 优 |
| 50th percentile service time | | ms | term | 2.49755 | 2.59268 | 0.09513 | 3.67% | 优 |
| 90th percentile service time | | ms | term | 2.71322 | 2.92963 | 0.21641 | 7.39% | 优 |
| 99th percentile service time | | ms | term | 3.20673 | 4.4033 | 1.19657 | 27.17% | 优 |
| 99.9th percentile service time | | ms | term | 5.17998 | 7.02644 | 1.84646 | 26.28% | 优 |
| 100th percentile service time | | ms | term | 6.95227 | 10.7671 | 3.81483 | 35.43% | 优 |
| error rate | | % | term | 0 | 0 | 0 | - | - |
词组查询（query） | Min Throughput | | ops/s | phrase | 110.05 | 110.05 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | phrase | 110.07 | 110.07 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | phrase | 110.12 | 110.11 | -0.01 | -0.01% | 优 |
| 50th percentile latency | | ms | phrase | 3.09905 | 3.19024 | 0.09119 | 2.86% | 优 |
| 90th percentile latency | | ms | phrase | 3.62549 | 3.74124 | 0.11575 | 3.09% | 优 |
| 99th percentile latency | | ms | phrase | 4.55457 | 7.78506 | 3.23049 | 41.50% | 优 |
| 99.9th percentile latency | | ms | phrase | 8.29519 | 20.249 | 11.95381 | 59.03% | 优 |
| 100th percentile latency | | ms | phrase | 9.39771 | 23.2441 | 13.84639 | 59.57% | 优 |
| 50th percentile service time | | ms | phrase | 2.38248 | 2.3894 | 0.00692 | 0.29% | 优 |
| 90th percentile service time | | ms | phrase | 2.77084 | 2.75176 | -0.01908 | -0.69% | 优 |
| 99th percentile service time | | ms | phrase | 3.75448 | 5.27172 | 1.51724 | 28.78% | 优 |
| 99.9th percentile service time | | ms | phrase | 7.5974 | 19.2863 | 11.6889 | 60.61% | 优 |
| 100th percentile service time | | ms | phrase | 8.98362 | 22.7147 | 13.73108 | 60.45% | 优 |
| error rate | | % | phrase | 0 | 0 | 0 | - | - |
不带缓存的聚合查询（aggregation） | Min Throughput | | ops/s | country_agg_uncached | 3.6 | 3.6 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | country_agg_uncached | 3.61 | 3.6 | -0.01 | -0.28% | 优 |
| Max Throughput | | ops/s | country_agg_uncached | 3.61 | 3.61 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | country_agg_uncached | 157.466 | 179.894 | 22.428 | 12.47% | 优 |
| 90th percentile latency | | ms | country_agg_uncached | 217.148 | 285.853 | 68.705 | 24.04% | 优 |
| 99th percentile latency | | ms | country_agg_uncached | 233.185 | 294.468 | 61.283 | 20.81% | 优 |
| 100th percentile latency | | ms | country_agg_uncached | 233.227 | 297.423 | 64.196 | 21.58% | 优 |
| 50th percentile service time | | ms | country_agg_uncached | 156.197 | 174.585 | 18.388 | 10.53% | 优 |
| 90th percentile service time | | ms | country_agg_uncached | 215.852 | 285.207 | 69.355 | 24.32% | 优 |
| 99th percentile service time | | ms | country_agg_uncached | 232.177 | 287.223 | 55.046 | 19.16% | 优 |
| 100th percentile service time | | ms | country_agg_uncached | 232.321 | 287.558 | 55.237 | 19.21% | 优 |
| error rate | | % | country_agg_uncached | 0 | 0 | 0 | - | - |
带缓存的聚合查询（aggregation） | Min Throughput | | ops/s | country_agg_cached | 100.03 | 100.04 | 0.01 | 0.01% | 优 |
| Median Throughput | | ops/s | country_agg_cached | 100.05 | 100.05 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | country_agg_cached | 100.08 | 100.08 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | country_agg_cached | 2.44457 | 2.43436 | -0.01021 | -0.42% | 优 |
| 90th percentile latency | | ms | country_agg_cached | 2.97922 | 2.93955 | -0.03967 | -1.35% | 劣 |
| 99th percentile latency | | ms | country_agg_cached | 3.96393 | 3.89118 | -0.07275 | -1.87% | 劣 |
| 99.9th percentile latency | | ms | country_agg_cached | 5.3294 | 5.35195 | 0.02255 | 0.42% | 优 |
| 100th percentile latency | | ms | country_agg_cached | 7.9529 | 5.4249 | -2.528 | -46.60% | 劣 |
| 50th percentile service time | | ms | country_agg_cached | 1.71924 | 1.66542 | -0.05382 | -3.23% | 劣 |
| 90th percentile service time | | ms | country_agg_cached | 1.97892 | 1.84812 | -0.1308 | -7.08% | 劣 |
| 99th percentile service time | | ms | country_agg_cached | 2.22611 | 2.17057 | -0.05554 | -2.56% | 劣 |
| 99.9th percentile service time | | ms | country_agg_cached | 5.0967 | 4.25527 | -0.84143 | -19.77% | 劣 |
| 100th percentile service time | | ms | country_agg_cached | 7.02246 | 4.95142 | -2.07104 | -41.83% | 劣 |
| error rate | | % | country_agg_cached | 0 | 0 | 0 | - | - |
分页拉取 | Min Throughput | | pages/s | scroll | 20.04 | 20.05 | 0.01 | 0.05% | 优 |
| Median Throughput | | pages/s | scroll | 20.04 | 20.06 | 0.02 | 0.10% | 优 |
| Max Throughput | | pages/s | scroll | 20.05 | 20.07 | 0.02 | 0.10% | 优 |
| 50th percentile latency | | ms | scroll | 576.675 | 598.124 | 21.449 | 3.59% | 优 |
| 90th percentile latency | | ms | scroll | 585.156 | 620.373 | 35.217 | 5.68% | 优 |
| 99th percentile latency | | ms | scroll | 598.95 | 646.07 | 47.12 | 7.29% | 优 |
| 100th percentile latency | | ms | scroll | 602.009 | 646.523 | 44.514 | 6.89% | 优 |
| 50th percentile service time | | ms | scroll | 575.118 | 596.707 | 21.589 | 3.62% | 优 |
| 90th percentile service time | | ms | scroll | 583.906 | 618.824 | 34.918 | 5.64% | 优 |
| 99th percentile service time | | ms | scroll | 597.482 | 644.695 | 47.213 | 7.32% | 优 |
| 100th percentile service time | | ms | scroll | 600.578 | 644.847 | 44.269 | 6.87% | 优 |
| error rate | | % | scroll | 0 | 0 | 0 | - | - |
脚本查询（使用 expression 脚本） | Min Throughput | | ops/s | expression | 2 | 2 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | expression | 2 | 2 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | expression | 2 | 2 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | expression | 299.685 | 360.09 | 60.405 | 16.77% | 优 |
| 90th percentile latency | | ms | expression | 416.613 | 491.791 | 75.178 | 15.29% | 优 |
| 99th percentile latency | | ms | expression | 465.776 | 500.174 | 34.398 | 6.88% | 优 |
| 100th percentile latency | | ms | expression | 468.083 | 500.209 | 32.126 | 6.42% | 优 |
| 50th percentile service time | | ms | expression | 298.594 | 359.072 | 60.478 | 16.84% | 优 |
| 90th percentile service time | | ms | expression | 415.045 | 489.912 | 74.867 | 15.28% | 优 |
| 99th percentile service time | | ms | expression | 464.598 | 499.031 | 34.433 | 6.90% | 优 |
| 100th percentile service time | | ms | expression | 467.106 | 499.083 | 31.977 | 6.41% | 优 |
| error rate | | % | expression | 0 | 0 | 0 | - | - |
脚本查询（使用 painless 静态脚本，不动态取字段值） | Min Throughput | | ops/s | painless_static | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | painless_static | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | painless_static | 1.5 | 1.5 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | painless_static | 383.485 | 389.354 | 5.869 | 1.51% | 优 |
| 90th percentile latency | | ms | painless_static | 514.495 | 641.987 | 127.492 | 19.86% | 优 |
| 99th percentile latency | | ms | painless_static | 561.342 | 644.179 | 82.837 | 12.86% | 优 |
| 100th percentile latency | | ms | painless_static | 568.066 | 646.037 | 77.971 | 12.07% | 优 |
| 50th percentile service time | | ms | painless_static | 382.158 | 388.178 | 6.02 | 1.55% | 优 |
| 90th percentile service time | | ms | painless_static | 513.202 | 640.374 | 127.172 | 19.86% | 优 |
| 99th percentile service time | | ms | painless_static | 560.61 | 642.894 | 82.284 | 12.80% | 优 |
| 100th percentile service time | | ms | painless_static | 567.419 | 644.877 | 77.458 | 12.01% | 优 |
| error rate | | % | painless_static | 0 | 0 | 0 | - | - |
脚本查询（使用 painless 静态脚本，动态获取字段值） | Min Throughput | | ops/s | painless_dynamic | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | painless_dynamic | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | painless_dynamic | 1.5 | 1.5 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | painless_dynamic | 377.278 | 393.005 | 15.727 | 4.00% | 优 |
| 90th percentile latency | | ms | painless_dynamic | 517.496 | 633.853 | 116.357 | 18.36% | 优 |
| 99th percentile latency | | ms | painless_dynamic | 576.697 | 653.933 | 77.236 | 11.81% | 优 |
| 100th percentile latency | | ms | painless_dynamic | 580.017 | 660.309 | 80.292 | 12.16% | 优 |
| 50th percentile service time | | ms | painless_dynamic | 376.339 | 391.545 | 15.206 | 3.88% | 优 |
| 90th percentile service time | | ms | painless_dynamic | 516.407 | 632.713 | 116.306 | 18.38% | 优 |
| 99th percentile service time | | ms | painless_dynamic | 575.714 | 652.838 | 77.124 | 11.81% | 优 |
| 100th percentile service time | | ms | painless_dynamic | 579.642 | 659.102 | 79.46 | 12.06% | 优 |
| error rate | | % | painless_dynamic | 0 | 0 | 0 | - | - |
地理范围查询（基于高斯衰减函数） | Min Throughput | | ops/s | decay_geo_gauss_function_score | 1 | 1 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | decay_geo_gauss_function_score | 1 | 1 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | decay_geo_gauss_function_score | 1 | 1 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | decay_geo_gauss_function_score | 348.531 | 388.606 | 40.075 | 10.31% | 优 |
| 90th percentile latency | | ms | decay_geo_gauss_function_score | 398.351 | 472.049 | 73.698 | 15.61% | 优 |
| 99th percentile latency | | ms | decay_geo_gauss_function_score | 411.483 | 492.926 | 81.443 | 16.52% | 优 |
| 100th percentile latency | | ms | decay_geo_gauss_function_score | 457.615 | 494.574 | 36.959 | 7.47% | 优 |
| 50th percentile service time | | ms | decay_geo_gauss_function_score | 346.881 | 386.701 | 39.82 | 10.30% | 优 |
| 90th percentile service time | | ms | decay_geo_gauss_function_score | 397.08 | 470.906 | 73.826 | 15.68% | 优 |
| 99th percentile service time | | ms | decay_geo_gauss_function_score | 410.421 | 490.485 | 80.064 | 16.32% | 优 |
| 100th percentile service time | | ms | decay_geo_gauss_function_score | 455.704 | 492.584 | 36.88 | 7.49% | 优 |
| error rate | | % | decay_geo_gauss_function_score | 0 | 0 | 0 | - | - |
地理范围查询（基于高斯衰减函数，且脚本动态获取字段值） | Min Throughput | | ops/s | decay_geo_gauss_script_score | 1 | 1 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | decay_geo_gauss_script_score | 1 | 1 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | decay_geo_gauss_script_score | 1 | 1 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | decay_geo_gauss_script_score | 368.275 | 430.786 | 62.511 | 14.51% | 优 |
| 90th percentile latency | | ms | decay_geo_gauss_script_score | 414.905 | 539.778 | 124.873 | 23.13% | 优 |
| 99th percentile latency | | ms | decay_geo_gauss_script_score | 468.888 | 543.645 | 74.757 | 13.75% | 优 |
| 100th percentile latency | | ms | decay_geo_gauss_script_score | 477.25 | 546.901 | 69.651 | 12.74% | 优 |
| 50th percentile service time | | ms | decay_geo_gauss_script_score | 366.945 | 429.517 | 62.572 | 14.57% | 优 |
| 90th percentile service time | | ms | decay_geo_gauss_script_score | 413.609 | 538.256 | 124.647 | 23.16% | 优 |
| 99th percentile service time | | ms | decay_geo_gauss_script_score | 467.627 | 542.623 | 74.996 | 13.82% | 优 |
| 100th percentile service time | | ms | decay_geo_gauss_script_score | 475.367 | 545.409 | 70.042 | 12.84% | 优 |
| error rate | | % | decay_geo_gauss_script_score | 0 | 0 | 0 | - | - |
自定义评分函数查询（基于字段值定义函数） | Min Throughput | | ops/s | field_value_function_score | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | field_value_function_score | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | field_value_function_score | 1.51 | 1.51 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | field_value_function_score | 139.661 | 162.157 | 22.496 | 13.87% | 优 |
| 90th percentile latency | | ms | field_value_function_score | 183.675 | 215.863 | 32.188 | 14.91% | 优 |
| 99th percentile latency | | ms | field_value_function_score | 197.653 | 221.584 | 23.931 | 10.80% | 优 |
| 100th percentile latency | | ms | field_value_function_score | 202.345 | 228.148 | 25.803 | 11.31% | 优 |
| 50th percentile service time | | ms | field_value_function_score | 138.423 | 159.991 | 21.568 | 13.48% | 优 |
| 90th percentile service time | | ms | field_value_function_score | 182.404 | 214.416 | 32.012 | 14.93% | 优 |
| 99th percentile service time | | ms | field_value_function_score | 196.734 | 220.581 | 23.847 | 10.81% | 优 |
| 100th percentile service time | | ms | field_value_function_score | 201.442 | 226.309 | 24.867 | 10.99% | 优 |
| error rate | | % | field_value_function_score | 0 | 0 | 0 | - | - |
自定义评分函数查询（通过脚本动态获取字段值计算评分） | Min Throughput | | ops/s | field_value_script_score | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | field_value_script_score | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | field_value_script_score | 1.51 | 1.51 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | field_value_script_score | 188.952 | 189.786 | 0.834 | 0.44% | 优 |
| 90th percentile latency | | ms | field_value_script_score | 264.095 | 313.248 | 49.153 | 15.69% | 优 |
| 99th percentile latency | | ms | field_value_script_score | 271.153 | 326.875 | 55.722 | 17.05% | 优 |
| 100th percentile latency | | ms | field_value_script_score | 271.901 | 338.193 | 66.292 | 19.60% | 优 |
| 50th percentile service time | | ms | field_value_script_score | 187.218 | 187.56 | 0.342 | 0.18% | 优 |
| 90th percentile service time | | ms | field_value_script_score | 263.207 | 311.552 | 48.345 | 15.52% | 优 |
| 99th percentile service time | | ms | field_value_script_score | 269.578 | 325.759 | 56.181 | 17.25% | 优 |
| 100th percentile service time | | ms | field_value_script_score | 270.138 | 336.648 | 66.51 | 19.76% | 优 |
| error rate | | % | field_value_script_score | 0 | 0 | 0 | - | - |
大量 terms 条件查询（query） | Min Throughput | | ops/s | large_terms | 1.1 | 1.1 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | large_terms | 1.1 | 1.1 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | large_terms | 1.1 | 1.1 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | large_terms | 265.007 | 835.444 | 570.437 | 68.28% | 优 |
| 90th percentile latency | | ms | large_terms | 296.009 | 1134.63 | 838.621 | 73.91% | 优 |
| 99th percentile latency | | ms | large_terms | 310.358 | 1323.69 | 1013.332 | 76.55% | 优 |
| 100th percentile latency | | ms | large_terms | 311.049 | 1360.9 | 1049.851 | 77.14% | 优 |
| 50th percentile service time | | ms | large_terms | 256.372 | 774.779 | 518.407 | 66.91% | 优 |
| 90th percentile service time | | ms | large_terms | 287.851 | 1022.85 | 734.999 | 71.86% | 优 |
| 99th percentile service time | | ms | large_terms | 301.827 | 1033.06 | 731.233 | 70.78% | 优 |
| 100th percentile service time | | ms | large_terms | 302.251 | 1033.73 | 731.479 | 70.76% | 优 |
| error rate | | % | large_terms | 0 | 0 | 0 | - | - |
大量 terms 条件过滤查询（query、filter） | Min Throughput | | ops/s | large_filtered_terms | 1.1 | 1.1 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | large_filtered_terms | 1.1 | 1.1 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | large_filtered_terms | 1.1 | 1.1 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | large_filtered_terms | 268.135 | 778.729 | 510.594 | 65.57% | 优 |
| 90th percentile latency | | ms | large_filtered_terms | 304.158 | 1007.17 | 703.012 | 69.80% | 优 |
| 99th percentile latency | | ms | large_filtered_terms | 351.209 | 1107.21 | 756.001 | 68.28% | 优 |
| 100th percentile latency | | ms | large_filtered_terms | 352.003 | 1137.25 | 785.247 | 69.05% | 优 |
| 50th percentile service time | | ms | large_filtered_terms | 259.546 | 695.247 | 435.701 | 62.67% | 优 |
| 90th percentile service time | | ms | large_filtered_terms | 295.721 | 997.742 | 702.021 | 70.36% | 优 |
| 99th percentile service time | | ms | large_filtered_terms | 342.342 | 1023.46 | 681.118 | 66.55% | 优 |
| 100th percentile service time | | ms | large_filtered_terms | 343.378 | 1026.96 | 683.582 | 66.56% | 优 |
| error rate | | % | large_filtered_terms | 0 | 0 | 0 | - | - |
大量条件取反查询（query、must not） | Min Throughput | | ops/s | large_prohibited_terms | 1.1 | 1.1 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | large_prohibited_terms | 1.1 | 1.1 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | large_prohibited_terms | 1.1 | 1.1 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | large_prohibited_terms | 270.041 | 828.126 | 558.085 | 67.39% | 优 |
| 90th percentile latency | | ms | large_prohibited_terms | 310.351 | 1120.94 | 810.589 | 72.31% | 优 |
| 99th percentile latency | | ms | large_prohibited_terms | 347.414 | 1294.23 | 946.816 | 73.16% | 优 |
| 100th percentile latency | | ms | large_prohibited_terms | 349.499 | 1371 | 1021.501 | 74.51% | 优 |
| 50th percentile service time | | ms | large_prohibited_terms | 261.734 | 728.821 | 467.087 | 64.09% | 优 |
| 90th percentile service time | | ms | large_prohibited_terms | 302.279 | 1012.35 | 710.071 | 70.14% | 优 |
| 99th percentile service time | | ms | large_prohibited_terms | 339.278 | 1032.58 | 693.302 | 67.14% | 优 |
| 100th percentile service time | | ms | large_prohibited_terms | 340.817 | 1034.51 | 693.693 | 67.06% | 优 |
| error rate | | % | large_prohibited_terms | 0 | 0 | 0 | - | - |
降序排序查询 | Min Throughput | | ops/s | desc_sort_population | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | desc_sort_population | 1.51 | 1.51 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | desc_sort_population | 1.51 | 1.51 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | desc_sort_population | 58.5828 | 65.9265 | 7.3437 | 11.14% | 优 |
| 90th percentile latency | | ms | desc_sort_population | 77.9981 | 118.956 | 40.9579 | 34.43% | 优 |
| 99th percentile latency | | ms | desc_sort_population | 80.8863 | 119.312 | 38.4257 | 32.21% | 优 |
| 100th percentile latency | | ms | desc_sort_population | 83.1661 | 119.441 | 36.2749 | 30.37% | 优 |
| 50th percentile service time | | ms | desc_sort_population | 57.1212 | 64.2672 | 7.146 | 11.12% | 优 |
| 90th percentile service time | | ms | desc_sort_population | 76.7082 | 117.437 | 40.7288 | 34.68% | 优 |
| 99th percentile service time | | ms | desc_sort_population | 79.2907 | 117.796 | 38.5053 | 32.69% | 优 |
| 100th percentile service time | | ms | desc_sort_population | 81.6364 | 117.817 | 36.1806 | 30.71% | 优 |
| error rate | | % | desc_sort_population | 0 | 0 | 0 | - | - |
升序排序查询 | Min Throughput | | ops/s | asc_sort_population | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | asc_sort_population | 1.51 | 1.51 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | asc_sort_population | 1.51 | 1.51 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | asc_sort_population | 62.4328 | 86.1481 | 23.7153 | 27.53% | 优 |
| 90th percentile latency | | ms | asc_sort_population | 79.8441 | 123.886 | 44.0419 | 35.55% | 优 |
| 99th percentile latency | | ms | asc_sort_population | 83.9411 | 124.563 | 40.6219 | 32.61% | 优 |
| 100th percentile latency | | ms | asc_sort_population | 84.3925 | 125.311 | 40.9185 | 32.65% | 优 |
| 50th percentile service time | | ms | asc_sort_population | 61.0637 | 84.976 | 23.9123 | 28.14% | 优 |
| 90th percentile service time | | ms | asc_sort_population | 78.4101 | 122.197 | 43.7869 | 35.83% | 优 |
| 99th percentile service time | | ms | asc_sort_population | 82.2652 | 123.488 | 41.2228 | 33.38% | 优 |
| 100th percentile service time | | ms | asc_sort_population | 82.5616 | 124.006 | 41.4444 | 33.42% | 优 |
| error rate | | % | asc_sort_population | 0 | 0 | 0 | - | - |
升序排序后 after 跳转查询 | Min Throughput | | ops/s | asc_sort_with_after_population | 1.5 | 1.5 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | asc_sort_with_after_population | 1.51 | 1.5 | -0.01 | -0.67% | 优 |
| Max Throughput | | ops/s | asc_sort_with_after_population | 1.51 | 1.51 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | asc_sort_with_after_population | 88.1871 | 99.9363 | 11.7492 | 11.76% | 优 |
| 90th percentile latency | | ms | asc_sort_with_after_population | 127.995 | 173.22 | 45.225 | 26.11% | 优 |
| 99th percentile latency | | ms | asc_sort_with_after_population | 131.171 | 174.543 | 43.372 | 24.85% | 优 |
| 100th percentile latency | | ms | asc_sort_with_after_population | 132.181 | 174.627 | 42.446 | 24.31% | 优 |
| 50th percentile service time | | ms | asc_sort_with_after_population | 87.132 | 98.2278 | 11.0958 | 11.30% | 优 |
| 90th percentile service time | | ms | asc_sort_with_after_population | 126.818 | 171.284 | 44.466 | 25.96% | 优 |
| 99th percentile service time | | ms | asc_sort_with_after_population | 129.453 | 171.658 | 42.205 | 24.59% | 优 |
| 100th percentile service time | | ms | asc_sort_with_after_population | 130.452 | 171.671 | 41.219 | 24.01% | 优 |
| error rate | | % | asc_sort_with_after_population | 0 | 0 | 0 | - | - |
高基字段降序排序查询（基于 DistanceFeatureQuery 快速取 topK） | Min Throughput | | ops/s | desc_sort_geonameid | 6.02 | 6.02 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | desc_sort_geonameid | 6.02 | 6.02 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | desc_sort_geonameid | 6.03 | 6.03 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | desc_sort_geonameid | 7.4659 | 7.55118 | 0.08528 | 1.13% | 优 |
| 90th percentile latency | | ms | desc_sort_geonameid | 8.26766 | 9.07559 | 0.80793 | 8.90% | 优 |
| 99th percentile latency | | ms | desc_sort_geonameid | 8.72369 | 9.69573 | 0.97204 | 10.03% | 优 |
| 100th percentile latency | | ms | desc_sort_geonameid | 8.79956 | 10.4637 | 1.66414 | 15.90% | 优 |
| 50th percentile service time | | ms | desc_sort_geonameid | 6.59986 | 6.5233 | -0.07656 | -1.17% | 劣 |
| 90th percentile service time | | ms | desc_sort_geonameid | 7.24539 | 7.85882 | 0.61343 | 7.81% | 优 |
| 99th percentile service time | | ms | desc_sort_geonameid | 7.57925 | 8.40503 | 0.82578 | 9.82% | 优 |
| 100th percentile service time | | ms | desc_sort_geonameid | 7.64471 | 9.40507 | 1.76036 | 18.72% | 优 |
| error rate | | % | desc_sort_geonameid | 0 | 0 | 0 | - | - |
高基字段降序排序 after 跳转查询 | Min Throughput | | ops/s | desc_sort_with_after_geonameid | 6.01 | 6 | -0.01 | -0.17% | 优 |
| Median Throughput | | ops/s | desc_sort_with_after_geonameid | 6.01 | 6.01 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | desc_sort_with_after_geonameid | 6.02 | 6.01 | -0.01 | -0.17% | 优 |
| 50th percentile latency | | ms | desc_sort_with_after_geonameid | 89.4587 | 107.925 | 18.4663 | 17.11% | 优 |
| 90th percentile latency | | ms | desc_sort_with_after_geonameid | 119.777 | 154.588 | 34.811 | 22.52% | 优 |
| 99th percentile latency | | ms | desc_sort_with_after_geonameid | 123.271 | 155.65 | 32.379 | 20.80% | 优 |
| 100th percentile latency | | ms | desc_sort_with_after_geonameid | 123.628 | 156.32 | 32.692 | 20.91% | 优 |
| 50th percentile service time | | ms | desc_sort_with_after_geonameid | 88.512 | 107.08 | 18.568 | 17.34% | 优 |
| 90th percentile service time | | ms | desc_sort_with_after_geonameid | 118.72 | 153.345 | 34.625 | 22.58% | 优 |
| 99th percentile service time | | ms | desc_sort_with_after_geonameid | 122.79 | 153.867 | 31.077 | 20.20% | 优 |
| 100th percentile service time | | ms | desc_sort_with_after_geonameid | 122.791 | 154.586 | 31.795 | 20.57% | 优 |
| error rate | | % | desc_sort_with_after_geonameid | 0 | 0 | 0 | - | - |
高基字段升序排序查询（基于 DistanceFeatureQuery 快速取 topK） | Min Throughput | | ops/s | asc_sort_geonameid | 6.02 | 6.02 | 0 | 0.00% | 优 |
| Median Throughput | | ops/s | asc_sort_geonameid | 6.02 | 6.02 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | asc_sort_geonameid | 6.03 | 6.03 | 0 | 0.00% | 优 |
| 50th percentile latency | | ms | asc_sort_geonameid | 5.80593 | 5.78184 | -0.02409 | -0.42% | 优 |
| 90th percentile latency | | ms | asc_sort_geonameid | 6.55438 | 6.60544 | 0.05106 | 0.77% | 优 |
| 99th percentile latency | | ms | asc_sort_geonameid | 7.36432 | 8.35744 | 0.99312 | 11.88% | 优 |
| 100th percentile latency | | ms | asc_sort_geonameid | 7.49672 | 30.8939 | 23.39718 | 75.73% | 优 |
| 50th percentile service time | | ms | asc_sort_geonameid | 4.91916 | 4.95407 | 0.03491 | 0.70% | 优 |
| 90th percentile service time | | ms | asc_sort_geonameid | 5.61126 | 5.2657 | -0.34556 | -6.56% | 劣 |
| 99th percentile service time | | ms | asc_sort_geonameid | 6.12285 | 7.42835 | 1.3055 | 17.57% | 优 |
| 100th percentile service time | | ms | asc_sort_geonameid | 6.51222 | 29.8436 | 23.33138 | 78.18% | 优 |
| error rate | | % | asc_sort_geonameid | 0 | 0 | 0 | - | - |
高基字段升序排序 after 跳转查询 | Min Throughput | | ops/s | asc_sort_with_after_geonameid | 6.01 | 6 | -0.01 | -0.17% | 优 |
| Median Throughput | | ops/s | asc_sort_with_after_geonameid | 6.01 | 6.01 | 0 | 0.00% | 优 |
| Max Throughput | | ops/s | asc_sort_with_after_geonameid | 6.01 | 6.02 | 0.01 | 0.17% | 优 |
| 50th percentile latency | | ms | asc_sort_with_after_geonameid | 70.994 | 102.943 | 31.949 | 31.04% | 优 |
| 90th percentile latency | | ms | asc_sort_with_after_geonameid | 104.817 | 137.463 | 32.646 | 23.75% | 优 |
| 99th percentile latency | | ms | asc_sort_with_after_geonameid | 108.797 | 139.088 | 30.291 | 21.78% | 优 |
| 100th percentile latency | | ms | asc_sort_with_after_geonameid | 108.929 | 140.114 | 31.185 | 22.26% | 优 |
| 50th percentile service time | | ms | asc_sort_with_after_geonameid | 69.7056 | 101.661 | 31.9554 | 31.43% | 优 |
| 90th percentile service time | | ms | asc_sort_with_after_geonameid | 103.875 | 136.324 | 32.449 | 23.80% | 优 |
| 99th percentile service time | | ms | asc_sort_with_after_geonameid | 107.828 | 138.331 | 30.503 | 22.05% | 优 |
| 100th percentile service time | | ms | asc_sort_with_after_geonameid | 108.539 | 139.012 | 30.473 | 21.92% | 优 |
| error rate | | % | asc_sort_with_after_geonameid | 0 | 0 | 0 | - | - |
