Introduction
Tulip is a Kotlin SDK to create load testing, stress testing, and performance regression testing benchmarks for software systems. Tulip is a robust solution for evaluating the performance and stability of web applications under varying conditions, ensuring they can handle high traffic and stress levels efficiently. Similar to well-known tools like JMeter, Gatling, and Locust, Tulip provides powerful capabilities to simulate real-world scenarios and generate comprehensive reports, helping developers identify and resolve potential bottlenecks.
Maven Central
<dependency>
<groupId>io.github.wfouche.tulip</groupId>
<artifactId>tulip-runtime</artifactId>
<version>2.1.7</version>
</dependency>implementation("io.github.wfouche.tulip:tulip-runtime:2.1.7")///usr/bin/env jbang "$0" "$@" ; exit $\?
//DEPS io.github.wfouche.tulip:tulip-runtime:2.1.7
//SOURCES FmpHttpUser.java
//JAVA 21
import io.github.wfouche.tulip.api.TulipApi;
public class App {
public static void main(String[] args) {
TulipApi.runTulip("benchmark_config.json");
}
}Program structure
A Tulip benchmark program consists of four parts:
-
Benchmark configuration file specifying benchmark scenarios to be tested
-
User class that implement actions to be benchmarked
-
App class that activates the Tulip runtime
-
Tulip runtime .jar file available from Maven Central
|
The Tulip runtime is written in Kotlin. Tulip benchmark programs (App and User classes) can be written in any Java Virtual Machine (JVM) programming language. The following languages are currently supported:
-
Java, Kotlin
-
Scala, Groovy
Coordinated Omission
Tulip resolves the infamous Coordination Omission problem. It reports internal wait-time separately from external service-time. This is a reporting capability unique to Tulip. The internal wait-time statistics is used to tune the Tulip runtime environment by changing the number of worker threads, or changing the amount of heap space assigned to the Java VM, altering the throughput-rate, or adjusting worker thread input queue sizes; all in an attempt to reduce the internal wait-time to a reasonable level so that it has a negligible impact on end-user response times.
Coordinated Omission is the unintended back pressure a system under test can apply to a load generation tool that prevents that tool from accurately recording user experience. Response time = wait time + service time.
Reporting
Tulip uses HdrHistogram to display Latency by Percentile Distribution graphs.
|
The latency graphs are contained in the detailed performance report which the Tulip runtime creates for a benchmark suite.
