1. Abstract
Performance testing is no longer a niche activity reserved for specialists. As modern applications become increasingly distributed, QA engineers and developers need practical skills to validate system performance, identify bottlenecks, and detect performance regressions before they reach production.
 
This full-day, hands-on workshop introduces modern performance testing using a fully open-source stack built around k6 OSS and Grafana OSS. Participants will learn how to design, execute, and analyze load tests while gaining practical experience with performance metrics, dashboards, and investigations.
 
Starting with a simple demo application, attendees will progressively build load tests with k6, explore different workload models, define performance assertions and pass/fail criteria, parameterize test data, structure a test suite integrated into a CI/CD pipeline, and generate tests with k6 Studio.
 
The workshop also demonstrates how observability improves performance testing. Participants will collect test results with Prometheus, visualize performance data in Grafana, and correlate load-test results with application and infrastructure metrics, logs, and traces to better understand system behavior under load and quickly identify the root causes of performance issues. From there, testing extends beyond HTTP services to infrastructure components such as databases and WebSocket connections, and into hybrid tests that correlate real browser performance with backend load.
 
Finally, attendees will explore how AI assistants can accelerate performance testing by helping generate test scenarios, create dashboards, interpret results, and investigate regressions, reducing manual effort throughout the testing process.
 
By the end of the workshop, participants will understand the fundamentals of performance testing, know how to design and execute realistic load tests with k6, analyze results using an open-source observability stack, and apply AI-assisted techniques to make performance testing faster and more effective.

2.Agenda
● Performance testing fundamentals
● Workshop setup: installing the demo application and k6
● Create your first load test with k6
● Model workloads using request-based rate
● Define assertions and pass/fail criteria
● Parameterize test data for realistic, per-user traffic
● Structure a test suite and integrate it into a CI/CD pipeline
● Generate tests and handle dynamic data with k6 Studio
● Store and visualize performance results in Grafana
● Observe the system under test: correlate metrics, logs, and traces to find root causes
● Test infrastructure beyond HTTP: databases and WebSocket connections, plus custom metrics
● Run hybrid performance tests correlating real browser experience with backend load
● Explore the k6 ecosystem and extensibility
● Use AI assistants to accelerate test creation, analysis, and investigations

3. Objective
By the end of this workshop, participants will:
● Gain a solid understanding of the fundamentals of modern performance testing.
● Learn a practical workflow for designing, executing, and evaluating load tests.
● Understand how observability complements performance testing to diagnose bottlenecks and regressions.
● Become familiar with an end-to-end, fully open-source performance testing stack built around k6, Prometheus, and Grafana.
● Leave with the confidence to apply these techniques and tools to their own applications and testing workflows.

4. Target audience and prerequizites
Software developers, QA engineers, DevOps engineers, and anyone interested in performance testing, observability, or the Grafana ecosystem. No prior experience with k6 or load testing is required, although basic programming knowledge is recommended.

5 .Technical requirements
Each participant needs their own laptop, with:
● Administrator/install rights, running Linux, macOS, or Windows.
● Reliable Wi-Fi/Internet access, both to clone the workshop repository and to pull the Docker images used throughout the day. Pulling everything ahead of time is strongly recommended, as pulling on venue Wi-Fi during the session is a common source of delay.
● Docker Desktop (https://docs.docker.com/get-started/get-docker/) (or an equivalent engine), with at least 4 GB of RAM available to Docker (8 GB+ total system RAM recommended). The local stack runs QuickPizza, PostgreSQL, k6, Prometheus, Grafana, Loki, Tempo, Pyroscope, and Grafana Alloy at the same time.
● Ports 3000, 3100, 3200, 3333, 3334, 4040, 4317, 4318, 5432, 9090, and 12345 free on localhost. These need to be available for the stack's containers.
● Git.
● The k6 OSS (https://grafana.com/docs/k6/latest/set-up/install-k6/) CLI.
● k6 Studio (https://grafana.com/docs/k6-studio/set-up/install/).
● A code editor/IDE for writing and editing k6 scripts (JavaScript).

A free Grafana Cloud account (https://grafana.com/auth/sign-up/create-user). Some exercises use Grafana Assistant, which requires a Grafana Cloud connection even though Grafana itself runs locally.