First time load testing, some basic queries

Grafana k6 OSS Support
1

We are testing a web application with secured access via a custom login. K6, Gafana and Influx db. The error count rises and the response time increases with increased VU as is to be expected. How do we gain insight into the cause and find a resolution is something not clear from the documentation. Metrics being mapped onto actionable causes, For instance we get a bunch of errors like

|INFO[0256] signalr/start Status code 400 
|INFO[0256] Response body null
|INFO[0256] Response time 449.2371 ms 
|INFO[0256] Error
|INFO[0256] Error code 1400
|INFO[0014] Status code 500

Error code description states generic error, status code error shows server error. Is there a way to get more verbose results ? Things like user machine running out of TCP slots leading to failed requests is obviously useless in a website test. Can anyone point me to some resource on how to translate metrics or errors into actionable causes, specifically for websites. Focus is on increasing success rates then increasing response times.

2

This is because there isn’t a one-size-fits-all solution here. k6 is the tool that can potentially expose performance and other issues in your website or webapp or API. How you can resolve these issues varies greatly and depends entirely on the system you’re testing. Performance problems in a PHP website on a shared hosting are different than issues with a Java backend API or issues from a Go DB-heavy API.

In your case, the slice of log you posted contains a HTTP 400 status code error (i.e. one of your requests probably wasn’t correct), and a 500 status code, which probably means something on your backend was broken for that specific request. k6 can’t possibly know what exactly that was, or how to suggest an “actionable cause” to address the issue.