Summary: | InfluxDB results does not match with Jmeter result due to aggregation | ||
---|---|---|---|
Product: | JMeter - Now in Github | Reporter: | Jyoti Prakash <jp.accr> |
Component: | Main | Assignee: | JMeter issues mailing list <issues> |
Status: | NEW --- | ||
Severity: | enhancement | Keywords: | PatchAvailable |
Priority: | P2 | ||
Version: | 5.3 | ||
Target Milestone: | JMETER_5.4 | ||
Hardware: | All | ||
OS: | All | ||
Attachments: | Example screenshot |
This is why I created and use https://github.com/apache/jmeter/pull/544 This issue has been migrated to GitHub: https://github.com/apache/jmeter/issues/5350 |
Created attachment 37335 [details] Example screenshot JMeter, by default adds the sampler metrics for a common sampler in the span of influxDB interval duration and send aggregated values to Influx DB after given interval. Jmeter calculates the transaction metrics (Min, Average, Maximum, 90% Line, etc) from raw results where all transaction metrics are stored vs while sending into InfluxDB from Jmeter, It aggregates metrics according to time interval instead of sending raw results. This causes the discrepancies between Influx DB metrics and Jmeter Metrics. If we can modify code to send all result samples rather that aggregation then it will resolve mismatch issue. Also, we can just send response time and need not to send min, max, and percentile because for 1 result row, time will be same for all computation, however for Summary we may need that.