The repo for SoCC23 Paper: LatenSeer: Causal Modeling of End-to-End Latency Distributions by Harnessing Distributed Tracing.
LatenSeer is a modeling framework for estimating end-to-end latency distributions in microservice-based web applications.
- An offline tool.
- Harnesses distributed tracing.
- Enables what-if analysis, predicting the potential impacts on end-to-end latency distribution due to various changes in service latencies.
The repo includes the source code of LatenSeer, scripts for setting up DSB experiments, and a script for collecting DSB Jaeger traces.
First, you can create a virtual environment and activate it:
python -m venv .venv
source .venv/bin/activate
Then install LatenSeer:
pip install .
Now you are good to do a test run. We provide a simple tutorial about how to use LatenSeer.
@inproceedings{zhang2023-latenseer,
title={LatenSeer: Causal Modeling of End-to-End Latency Distributions by Harnessing Distributed Tracing},
author={Zhang, Yazhuo and Isaacs, Rebecca and Yue, Yao and Yang, Juncheng and Zhang, Lei and Vigfusson, Ymir},
booktitle={ACM Symposium on Cloud Computing (SoCC’23)},
year={2023}
}