Performance load profile generation

2. STCLab LoadTester

Captures and anonymizes real production HTTP traffic, then replays it in test environments. It supports dynamic parameter substitution and realistic user journeys.www.stclab.com

Best for: Reproducing real traffic patterns instead of manually inventing user behavior.

Website

3. Tricentis NeoLoad

Combines AI-assisted performance engineering with recorded user journeys, configurable load profiles, and integrations with APM tools such as Datadog and Dynatrace.www.tricentis.com+2

Best for: Enterprise teams that need broad application coverage, CI/CD integration, and performance analysis.

Website

Why your tests may be inconsistent

Whichever tool you choose, look for these capabilities:

• Production-derived traffic mix: Replicates the proportions of reads, writes, searches, logins, and other user journeys.
• Realistic arrival rates: Models bursts, quiet periods, peak hours, and ramp-up patterns rather than simply maintaining a fixed number of virtual users.
• Repeatable test profiles: Saves a traffic profile so you can rerun equivalent tests across builds.
• Safe traffic replay: Masks sensitive production data and handles session tokens, dynamic IDs, and test data correctly.
• Comparable results: Tracks throughput, p95/p99 latency, error rates, and resource utilization against a consistent baseline.

One important distinction: generating realistic traffic and reproducing the exact same traffic on every run are separate requirements. For your inconsistency problem, you need both realistic workload modeling and deterministic, repeatable test execution.

My recommendation

Start with Loadmill if you want AI-generated scenarios, and compare it with STCLab LoadTester if your priority is replaying real production traffic.

One question that would help narrow this down: What do you currently use for performance testing?

Current testing tool

Apache JMeter

Grafana k6

Gatling

LoadRunner / NeoLoad

Custom scripts or another tool

Where do your production patterns come from?

APM or observability tools

HTTP access logs or traces

Captured live traffic

We don't have a usable traffic dataset yet

Primary goal

Generate realistic load profiles automatically

Replay production traffic in a test environment

Both, with repeatable regression tests

[Find the best fit]

Top tools designed for this exact use case include:

• Loadmill : An AI-native testing platform that connects directly to your data sources (like proxy captures, API gateways, or HAR files) to automatically generate realistic, production-ready user scenarios and load profiles without manual scripting. It handles dynamic data correlation and mimics actual user behavior paths out of the box.[loadmill.com](https://www.loadmill.com/performance-testing)[[2](https://www.loadmill.com/performance-testing)](https://www.loadmill.com/blog/generative-ai-for-enterprise-testing-load-testing-in-the-age-of-ai)
• StormForge : Uses machine learning to analyze application and Kubernetes metrics, modeling complex production traffic patterns to simulate real-world spikes and steady states. It is particularly strong if your services run in containerized environments and you want to tie load testing directly to resource optimization.[loadmill.com](https://www.testmuai.com/blog/ai-in-performance-testing/)[[2](https://www.loadmill.com/performance-testing)](https://www.cloudbolt.io/videos/stormforge-application-performance-optimization-demo/)

If you'd like to narrow this down, tell me:

• What format is your production traffic data currently in (e.g., NGINX/Access logs, HAR files, APM traces)?
• Are your services hosted in Kubernetes/Cloud or a traditional server environment?

I can help you figure out which tool integrates best with your current stack.