@@ -10,6 +10,8 @@ This directory contains comprehensive testing tools for validating the SLA plann
The SLA planner monitors metrics every 60 seconds (default adjustment interval) and scales
prefill/decode workers based on TTFT, ITL, and request patterns.
To setup the environment, simply use the released docker images for any backends, or build your own docker image following the READMEs in `./components/backends/<vllm/sglang/trtllm>/README.md`, or follow the `Developing Locally` section in [README.md](../../README.md) to setup the environment locally.
## Pre-Requisite: Pre-Deployment Profiling Data
You have two options to obtain the pre-deployment profiling data: