Skip to main content
The Latency evaluator measures the round-trip time from Adaline to the LLM provider. The steps below describe how to set up the Latency evaluator:
1

Select the Latency evaluator

Add the Latency evaluator
2

Link a dataset

Give a name to the evaluator, link the dataset to it, and select the latency threshold.Selecting the latency thresholdChoose among the following:
  • less than: The response must be faster than your threshold.
  • greater than: The response must be slower than your threshold.
  • equal to: The response must match your exact timing requirement.
3

Define the unit of measure

Define the unit of measure
4

Execute the evaluation and see the results

Click on Evaluate to run the evaluation and see the results.Visualize the results of the latency evaluator
Note on Prompt Chaining: When your prompt uses prompt variables (child prompts), latency is calculated based on the slowest execution at each level of the prompt chain. Think of it like a graph: prompts at the same depth level execute in parallel, and we take the maximum time from that level. The total latency is the sum of the maximum latencies from each level. For example, if Prompt A calls Prompts B and C in parallel (level 1), and Prompt C calls Prompts D and E in parallel (level 2), the latency would be: max(B, C) + max(D, E) + A’s own latency.