{"id":6934,"date":"2019-10-04T16:50:48","date_gmt":"2019-10-04T20:50:48","guid":{"rendered":"https:\/\/www.redline13.com\/blog\/?p=6934"},"modified":"2019-10-04T16:50:48","modified_gmt":"2019-10-04T20:50:48","slug":"advanced-load-testing-troubleshooting-the-case-of-the-last-minute-spike","status":"publish","type":"post","link":"https:\/\/www.redline13.com\/blog\/2019\/10\/advanced-load-testing-troubleshooting-the-case-of-the-last-minute-spike\/","title":{"rendered":"Advanced Load Testing Troubleshooting – the Case of the Last Minute Spike"},"content":{"rendered":"

Anyone who has done load testing knows it\u2019s a lot more than creating a script and running it. There are going to be times when it requires what we call \u2018Advanced Load Testing Troubleshooting\u2019.<\/p>\n

Here is a real case of a RedLine13 customer who was trying to figure out what was causing an unusual load testing graph result. The issue was the graph Percentiles per Page Response Time graph. But sometimes the problem we see is not the real problem. Let’s not get ahead of ourselves. So let the troubleshooting begin.<\/p>\n

The (Perceived) Problem<\/h3>\n

For all the tests the company was running, they saw the behavior that at the end of the test at exactly the last 1 minute with a sudden spike in the graph.<\/p>\n

<\/p>\n

Their understanding of the above graph behavior is that, since they stop the test suddenly at the end of the test, there may be some requests being sent but did not receive the response due to that spike is happening. Their questions:<\/p>\n

    \n
  1. Would this sudden spike cause any effect on the overall tests response times (they consider 90 Percentile values)?<\/li>\n
  2. Is there a way they could get response times table values by eliminating the last 5 minutes?<\/li>\n<\/ol>\n

    Our initial response was that they could do this manually:<\/p>\n