SleeperGenerator

Mock a Generator component for benchmarks and load testing. This component sleeps for a configurable duration and returns a placeholder answer.

Key Features

• Simulates Generator latency without calling an external LLM API.
• Sleep duration follows a normal distribution around a configurable mean.
• Returns a configurable placeholder answer and metadata.
• Useful for pipeline performance testing and benchmarking.

Configuration

1. Drag the SleeperGenerator component onto the canvas from the Component Library.
2. Click the component to open the configuration panel.
3. Set mean_sleep_in_seconds, sleep_scale, answer, and optional meta values.

Connections

SleeperGenerator accepts a prompt string as input. It outputs replies — a list containing the placeholder answer — and meta — optional metadata.

Connect a PromptBuilder to the prompt input. Connect the replies output to DeepsetAnswerBuilder or a pipeline output, the same way you would connect a real Generator.

Usage Examples

Basic Configuration

  SleeperGenerator:
    type: deepset_cloud_custom_nodes.generators.sleeper.SleeperGenerator
    init_parameters:
      mean_sleep_in_seconds: 5
      sleep_scale: 1.0
      answer: Placeholder response
      meta: {}

Using the Component in a Pipeline

This example replaces a real Generator with SleeperGenerator for load testing:

# haystack-pipeline
components:
  sleeper:
    type: deepset_cloud_custom_nodes.generators.sleeper.SleeperGenerator
    init_parameters:
      mean_sleep_in_seconds: 5
      sleep_scale: 1.0
      answer: Placeholder response
      meta: {}

inputs:
  prompt:
    - sleeper.prompt

outputs:
  replies: sleeper.replies

Parameters

Inputs

Parameter	Type	Default	Description
prompt	str		Prompt input. The component does not use this value but accepts it to match the Generator interface.
generation_kwargs	Optional[Dict[str, Any]]	None	Generation kwargs. The component does not use this value but accepts it to match the Generator interface.

Outputs

Parameter	Type	Default	Description
replies	List[str]		List containing the placeholder answer.
meta	List[Dict[str, Any]]		Metadata returned with the answer.

Init Parameters

These are the parameters you can configure in Pipeline Builder:

Parameter	Type	Default	Description
mean_sleep_in_seconds	float	10	Mean sleep duration in seconds before returning the answer.
sleep_scale	float	1.0	Standard deviation for the sleep duration.
answer	str	Placeholder	Text returned as the Generator reply.
meta	Optional[Dict[str, Any]]	None	Metadata returned with the answer.

Run Method Parameters

These are the parameters you can configure for the component's run() method. This means you can pass these parameters at query time through the API, in Playground, or when running a job. For details, see Modify Pipeline Parameters at Query Time.

Parameter	Type	Default	Description
prompt	str		Prompt input.
generation_kwargs	Optional[Dict[str, Any]]	None	Generation kwargs.