我有一个JMeter测试，当我运行此测试时。
测试主要包括GET请求
我在这里进行了两次测试：
Waiting for possible Shutdown/StopTestNow/Heapdump message on port 4445
summary + 1 in 00:00:06 = 0.2/s Avg: 4601 Min: 4601 Max: 4601 Err: 0 (0.00%) Active: 750 Started: 750 Finished: 0
summary + 18 in 00:00:25 = 0.7/s Avg: 9483 Min: 4525 Max: 29202 Err: 0 (0.00%) Active: 750 Started: 750 Finished: 0
summary = 19 in 00:00:31 = 0.6/s Avg: 9226 Min: 4525 Max: 29202 Err: 0 (0.00%)
summary + 551 in 00:00:30 = 18.1/s Avg: 11589 Min: 67 Max: 59660 Err: 392 (71.14%) Active: 750 Started: 750 Finished: 0
summary = 570 in 00:01:01 = 9.3/s Avg: 11511 Min: 67 Max: 59660 Err: 392 (68.77%)
summary + 189 in 00:00:30 = 6.4/s Avg: 52975 Min: 70 Max: 88945 Err: 42 (22.22%) Active: 750 Started: 750 Finished: 0
summary = 759 in 00:01:31 = 8.4/s Avg: 21836 Min: 67 Max: 88945 Err: 434 (57.18%)
summary + 218 in 00:00:30 = 7.2/s Avg: 95267 Min: 453 Max: 119995 Err: 0 (0.00%) Active: 750 Started: 750 Finished: 0
summary = 977 in 00:02:01 = 8.1/s Avg: 38221 Min: 67 Max: 119995 Err: 434 (44.42%)
summary + 194 in 00:00:29 = 6.6/s Avg: 126743 Min: 952 Max: 149166 Err: 0 (0.00%) Active: 750 Started: 750 Finished: 0
summary = 1171 in 00:02:31 = 7.8/s Avg: 52886 Min: 67 Max: 149166 Err: 434 (37.06%)
summary + 209 in 00:00:31 = 6.7/s Avg: 123485 Min: 885 Max: 165675 Err: 7 (3.35%) Active: 750 Started: 750 Finished: 0
summary = 1380 in 00:03:02 = 7.6/s Avg: 63579 Min: 67 Max: 165675 Err: 441 (31.96%)
summary + 227 in 00:00:30 = 7.7/s Avg: 113494 Min: 888 Max: 202528 Err: 6 (2.64%) Active: 750 Started: 750 Finished: 0
summary = 1607 in 00:03:31 = 7.6/s Avg: 70629 Min: 67 Max: 202528 Err: 447 (27.82%)
summary + 264 in 00:00:30 = 8.8/s Avg: 100507 Min: 960 Max: 181355 Err: 0 (0.00%) Active: 750 Started: 750 Finished: 0
summary = 1871 in 00:04:01 = 7.8/s Avg: 74845 Min: 67 Max: 202528 Err: 447 (23.89%)
summary + 218 in 00:00:30 = 7.3/s Avg: 97068 Min: 909 Max: 156728 Err: 5 (2.29%) Active: 750 Started: 750 Finished: 0
summary = 2089 in 00:04:31 = 7.7/s Avg: 77164 Min: 67 Max: 202528 Err: 452 (21.64%)
summary + 208 in 00:00:29 = 7.1/s Avg: 94552 Min: 650 Max: 143990 Err: 4 (1.92%) Active: 750 Started: 750 Finished: 0
summary = 2297 in 00:05:01 = 7.6/s Avg: 78739 Min: 67 Max: 202528 Err: 456 (19.85%)
summary + 201 in 00:00:32 = 6.2/s Avg: 107602 Min: 81760 Max: 151824 Err: 0 (0.00%) Active: 548 Started: 750 Finished: 202
summary = 2498 in 00:05:33 = 7.5/s Avg: 81061 Min: 67 Max: 202528 Err: 456 (18.25%)
summary + 210 in 00:00:28 = 7.5/s Avg: 112283 Min: 83132 Max: 150911 Err: 3 (1.43%) Active: 339 Started: 750 Finished: 411
summary = 2708 in 00:06:01 = 7.5/s Avg: 83483 Min: 67 Max: 202528 Err: 459 (16.95%)
summary + 224 in 00:00:31 = 7.3/s Avg: 105992 Min: 61142 Max: 148336 Err: 5 (2.23%) Active: 115 Started: 750 Finished: 635
summary = 2932 in 00:06:32 = 7.5/s Avg: 85202 Min: 67 Max: 202528 Err: 464 (15.83%)
summary + 114 in 00:00:14 = 8.3/s Avg: 124572 Min: 96381 Max: 153037 Err: 2 (1.75%) Active: 0 Started: 750 Finished: 750
summary = 3046 in 00:06:46 = 7.5/s Avg: 86676 Min: 67 Max: 202528 Err: 466 (15.30%)
Tidying up ... @ Tue Sep 19 13:54:15 BST 2017 (1505825655017)
第二个：
Waiting for possible Shutdown/StopTestNow/Heapdump message on port 4445
summary + 475 in 00:01:51 = 4.3/s Avg: 72697 Min: 2707 Max: 106641 Err: 0 (0.00%) Active: 750 Started: 750 Finished: 0
summary + 140 in 00:00:35 = 4.0/s Avg: 123952 Min: 106596 Max: 145691 Err: 0 (0.00%) Active: 750 Started: 750 Finished: 0
summary = 615 in 00:02:26 = 4.2/s Avg: 84365 Min: 2707 Max: 145691 Err: 0 (0.00%)
summary + 240 in 00:00:29 = 8.4/s Avg: 136274 Min: 107411 Max: 164554 Err: 0 (0.00%) Active: 750 Started: 750 Finished: 0
summary = 855 in 00:02:55 = 4.9/s Avg: 98936 Min: 2707 Max: 164554 Err: 0 (0.00%)
summary + 111 in 00:01:30 = 1.2/s Avg: 66969 Min: 3932 Max: 197338 Err: 51 (45.95%) Active: 750 Started: 750 Finished: 0
summary = 966 in 00:04:25 = 3.6/s Avg: 95262 Min: 2707 Max: 197338 Err: 51 (5.28%)
summary + 190 in 00:00:27 = 6.9/s Avg: 197033 Min: 179995 Max: 219692 Err: 0 (0.00%) Active: 750 Started: 750 Finished: 0
summary = 1156 in 00:04:53 = 4.0/s Avg: 111989 Min: 2707 Max: 219692 Err: 51 (4.41%)
summary + 727 in 00:00:28 = 25.5/s Avg: 61250 Min: 66 Max: 205546 Err: 487 (66.99%) Active: 610 Started: 750 Finished: 140
summary = 1883 in 00:05:21 = 5.9/s Avg: 92400 Min: 66 Max: 219692 Err: 538 (28.57%)
summary + 250 in 00:00:30 = 8.3/s Avg: 175360 Min: 81942 Max: 193665 Err: 0 (0.00%) Active: 360 Started: 750 Finished: 390
summary = 2133 in 00:05:51 = 6.1/s Avg: 102123 Min: 66 Max: 219692 Err: 538 (25.22%)
summary + 359 in 00:00:19 = 19.3/s Avg: 102742 Min: 68901 Max: 192077 Err: 0 (0.00%) Active: 0 Started: 750 Finished: 750
summary = 2492 in 00:06:10 = 6.7/s Avg: 102212 Min: 66 Max: 219692 Err: 538 (21.59%)
Tidying up ... @ Tue Sep 19 14:01:48 BST 2017 (1505826108986)
两者之间没有更改任何设置。但是为什么我在第一次运行时会再增加三个样品呢？在JMeter 3.1上针对EC2实例运行此命令。没有CPU或内存峰值，NGINX日志看起来也不错。此外，还有duraiton和线程。
答案 0 :(得分：0)
调查平均响应时间：
86676 102212 因此，由于响应时间增加，您执行的样本较少。
如果您的测试应用程序看起来很好，瓶颈可能在 JMeter端，如果JMeter无法足够快地发送请求 - 响应时间会更长。所以我建议在JMeter端设置CPU，RAM，Swap等的监控。您可以使用JMeter PerfMon Plugin。
另请确保您遵循9 Easy Solutions for a JMeter Load Test “Out of Memory” Failure指南中的建议。
答案 1 :(得分：0)
在查看响应时间之前，我会开始担心错误率。
如此高的错误率，请首先检查您的脚本，看看您是否有可变的（相关的）必需ID，并正确地再现了访问。
然后缓慢增加100，然后增加200 ....并检查会发生什么。