Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/gatling/gatling.io-doc/llms.txt

Use this file to discover all available pages before exploring further.

Trends and run comparison give you the longitudinal view of your application’s performance that a single run report cannot provide. While a report answers “how did this run perform?”, trends answer “is performance improving or degrading over time?”. Gatling Enterprise Edition captures run results after each execution and surfaces them through two complementary features: Run Trends, which gives a high-level bar-chart overview of your most recent runs, and Run Comparison, which lets you select up to five runs and compare them side by side across eleven different metrics at request-level granularity.
Run trends and run comparison are only available on Gatling Enterprise paid plans.
  • Run Trends: click a test name in the tests table to open its run trends page.
  • Run Comparison: click the Compare runs button on the run trends page or on any individual run’s report page.
The run trends page shows bar charts for your most recent 10 successfully completed runs. Runs that failed or were aborted are excluded from the charts. Each run is identified by its sequential run number on the x-axis; hover over a bar to see the run title and value.

Displayed metrics

Bars divided between OK (successful responses) and KO (error responses).Customisation options:
  • Toggle between absolute counts and percentages using the Show values / Show % toggle in the top-right corner of the graph
  • Filter to show only OK, only KO, or both by clicking the legend labels beneath the graph
Click on a run in the left-side menu to navigate to that run’s full interactive report.
Runs default to sequential numbers (Run 1, Run 2, etc.). To add a meaningful title:
  1. Select the run in the left-side menu
  2. Click the edit icon next to the run title (shown as Untitled run if not previously named)
  3. Enter a title and click Save
  4. Click Trends in the left-side menu to return to the trends view
Remove a run that contains no useful data (e.g., a run where the build step failed before any requests were sent):
  1. Select the run in the left-side menu
  2. Click the three-dot dropdown menu on the right side of the run summary
  3. Select Delete run
  4. Click Confirm in the confirmation modal

Run Comparison

Run comparison lets you select up to 5 runs and compare a single performance metric across all of them simultaneously, with the ability to adjust the analysis time window and designate a reference run for percentage-deviation calculations.

Selecting runs

1

Enter compare mode

Click Compare runs on the trends page or on any run’s report page.
2

Choose runs

In the left-side menu, click the empty checkbox next to each run you want to compare. You can select up to 5 runs.
3

Deselect runs

Click a filled checkbox or click the X next to the run number under Selected runs to remove a run from the comparison.

Selecting a metric

Use the dropdown in the Compare runs panel to change the metric being compared. The chart and table update dynamically. Available metrics:
MetricDescription
Concurrent usersNumber of simultaneous virtual users (default)
CPU usageLoad generator CPU utilisation percentage
ErrorsTotal error count
Response time p50Median response time
Response time p9090th percentile response time
Response time p9595th percentile response time
Response time p9999th percentile response time
Response time p99.999.9th percentile response time
Response time p99.9999.99th percentile response time
Response time maximumMaximum observed response time
ThroughputRequests per second

Adjusting the time window

By default, the entire run duration is compared. To zoom in on a specific time window:
  • Drag the left and right sliders to the desired boundaries
  • Drag and drop directly on the chart to highlight a region
  • Double-click anywhere on the graph to reset to the full run duration
Narrowing the time window is useful for isolating a specific load phase—such as the peak plateau or the ramp-down—for granular comparison.

Setting a reference run

The comparison table shows each metric value alongside a percent deviation from the reference run. The first selected run is the default reference. To change the reference run:
  1. Click the pin icon (📍) next to a different run in the comparison table
  2. The selected run moves to the left column and becomes the new reference
  3. All deviation percentages recalculate relative to the new reference
Pin your known-good baseline run as the reference to immediately see how newer runs deviate. A positive deviation on response time (slower) or errors (more) is a regression signal; a negative deviation is an improvement.

Regression detection workflow

1

Establish a baseline

After a stable, representative run, note the run number and save its title as “Baseline — YYYY-MM-DD”.
2

Monitor run trends

After each CI pipeline run, check the Response Time Percentiles and Response Status charts on the trends page. A consistent upward trend in p99 or a rising KO bar is an early regression signal.
3

Drill down with run comparison

Select the baseline run and the suspicious run in comparison mode. Pin the baseline as the reference and switch through metrics to identify which request or phase is responsible for the regression.
4

Investigate the report

Click through to the full report for the regressing run and use the Requests panel with time-window selection to pinpoint when during the run the degradation began.

Build docs developers (and LLMs) love