/benchmark

description: 'Design and coordinate a reproducible Java performance test.'
argument-hint: '[target]'
model: 'inherit'
agent: 'plinth-java-performance'
tools:
  - 'Read'
  - 'Write'
  - 'Edit'
  - 'Bash'

benchmark

Select and coordinate an appropriate Java performance test with reproducible workload, environment, thresholds, and result artifacts.

Usage

/benchmark <target> [objective-or-threshold] [workload] [environment] [preferred-tool]

Accepted inputs

• Target application, endpoint, workflow, method, or component
• Performance objective or threshold
• Expected workload and environment
• Optional preferred tool

Owner and skills

• Owner: `@plinth-java-performance`
• Associated skills: `@151-java-performance-jmeter`, `@152-java-performance-gatling`, and existing Maven/JMH guidance

Tool selection

| Need | Tool | | --- | --- | | HTTP/API load and performance test | JMeter or Gatling | | Scenario-oriented load model and reports | Gatling | | Isolated JVM method or component microbenchmark | JMH |

Workflow

1. Clarify target boundary, objective, thresholds, workload, and environment. 2. Select JMeter, Gatling, or JMH and record the rationale. 3. Define warm-up, duration, concurrency, data setup, and result artifacts. 4. Generate or coordinate the reproducible performance workflow using the selected skill or Maven/JMH guidance. 5. Evaluate results against explicit thresholds. 6. Report limitations, environment metadata, and whether results are comparable.

Output

• Selected tool and rationale
• Reproducible test configuration
• Baseline or result artifacts
• Threshold assessment
• Limitations and environment metadata

Safeguards

• Do not present non-equivalent runs as valid before/after comparisons.
• Do not use JMeter or Gatling for isolated JVM microbenchmarks when JMH is the correct boundary.
• Do not use JMH for end-to-end load behavior.
• Keep workload, environment, and threshold assumptions explicit.