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.

Interpreting load test results has traditionally required deep performance engineering expertise. A p99 latency spike might be a meaningful regression or statistical noise; a rising error rate might indicate an application bug or a saturated database connection pool. Gatling Enterprise Edition’s AI analysis features remove that friction by generating structured, plain-language summaries directly on your run results pages — so engineers of all experience levels can quickly understand what happened, compare runs across deploys, and identify drifting trends before they become incidents.
AI analysis features are available on Gatling Enterprise Edition. They can be disabled organisation-wide by an administrator. If you do not see the AI buttons described in this guide, contact your organisation admin.

AI Run Summary

After a simulation finishes, the AI Run Summary generates an instant plain-language analysis of that run’s results — without you having to interpret percentile charts or error distribution tables manually.

How to Access

1

Open a completed run

From the Runs list in Gatling Enterprise Edition, click on any completed run to open its detail page.
2

Click AI Run Summary

On the run results page, click the AI Run Summary button. The summary appears inline — no page reload, no separate view.

What the Summary Contains

The report is structured into two sections: Summary — A short, plain-language overview of the run as a whole: whether load was applied as intended, the overall error picture, and the most notable observation. Insights — A breakdown by area, shown only when relevant data exists:
Insight areaWhat it covers
Response timesAnalysis of latency distribution and key percentiles (p50, p95, p99)
Injection profileWhether virtual users were injected as configured
ErrorsObservations on error rates and error types, if any errors occurred
AssertionsInterpretation of pass/fail assertion results, if assertions were defined
AI Run Summary is generated on demand — it is not computed automatically when a run completes. Each click generates a fresh analysis; results may vary slightly between generations on the same data.

Use Case

Use AI Run Summary after every scheduled run to quickly triage whether the result needs deeper investigation. Instead of opening and reading raw charts for routine runs, let the summary tell you if everything looks nominal — and flag the ones that don’t.

AI Run Comparison

The AI Run Comparison feature analyses two to five runs of the same simulation side-by-side and produces a structured breakdown of what changed, what regressed, and what to investigate next.

How to Access

1

Open a run detail page

Navigate to any completed run’s detail page in Gatling Enterprise Edition.
2

Click Compare runs

Click the Compare runs button to open the Run Comparison view.
3

Select runs to compare

Select between 2 and 5 runs from the list. The Compare with AI button becomes active once at least 2 runs are selected.
4

Click Compare with AI

The AI report appears above the comparison chart. The chart remains fully interactive for manual exploration.

What the Report Contains

The comparison report has three structured blocks: Findings — Opens with a one-sentence summary of the most significant pattern, followed by a detailed breakdown: throughput changes, error count differences, latency percentile shifts, CPU usage deltas, and request-level highlights where relevant. Recommendations — Opens with a one-sentence focus for your investigation, followed by 2–4 concrete actions to take before the next run. Explore in the chart — Points you to specific metrics and run pairs to examine in the interactive chart below the report. The report also includes:
TagDescription
VerdictSimilar, SomeDiscrepancies, or Divergent — a one-word summary of how the runs relate
Confidence levelLow, Medium, or High — reflects how much signal was available. A low confidence often means a run was stopped early or had insufficient data
Selecting the same set of runs again retrieves the previously generated report rather than generating a new one.

Use Case

Use AI Run Comparison whenever you deploy a change that might affect performance: a new service version, a dependency upgrade, an infrastructure scaling event, or a configuration change. Select the last run before the change and the first run after it, and let the AI tell you whether the change introduced a regression. AI Trends Analysis reads your last 10 simulation runs and produces a structured health report covering what is stable, what is drifting, and what action to take before the next run.

How to Access

1

Open a test

Navigate to a simulation (test) in Gatling Enterprise Edition.
2

Select Trends

Click Trends in the left navigation panel to open the trends view.
3

Click Analyze with AI

In the AI Trends Analysis block, click Analyze with AI. The report appears above the trends charts.
Once generated, the AI Trends Analysis report persists on the trends page. The Analyze with AI button only becomes available again after new runs are added, at which point the existing report is marked as outdated.

What the Report Contains

Health Findings — Opens with the most significant health signal across the 10-run window, followed by observations about outlier runs: error spikes, tail latency anomalies, or runs that behave differently from the majority. Insights — Contextualises the findings: whether an anomaly is isolated or recurring, whether a spike indicates a defect burst or a structural performance issue developing over time. Recommendations — 2–4 concrete steps to apply before the next run: assertions to add, metrics to monitor more closely, or configuration to stabilise. The report also includes:
TagDescription
VerdictStable, SomeIssues, or Degrading — the overall health of the test’s trend
Confidence levelLow, Medium, or High — reflects the amount of signal available from the 10-run window

Use Case

Run AI Trends Analysis on a regular cadence — weekly or after every sprint — to catch gradual performance drift before it becomes a production incident. A Degrading verdict with High confidence is an early warning that warrants immediate investigation.

Summary: Which AI Feature to Use

AI Run Summary

Use after every run to instantly triage whether results need deeper investigation. Best for routine health checks.

AI Run Comparison

Use when you deploy a change and want to know whether performance regressed. Select the before and after runs for a structured diff.

AI Trends Analysis

Use on a regular schedule to spot gradual degradation across many runs. Covers the last 10 runs automatically.
AI-generated insights are provided for informational purposes only. Always verify AI-identified issues by examining the underlying metrics before making infrastructure or code changes based solely on AI output.

Build docs developers (and LLMs) love