Stopping Performance Regression via Changepoint Detection

Thursday, March 27, 2025 - 9:00 am9:20 am PDT
Authors: 

Joseph Cirella and Shanthini Velan, Bloomberg

Abstract: 

Bloomberg's Ticker Plant infrastructure provides real-time market data to almost all internal and external clients; any increase in latency impacts much of the company's real-time products. This talk discusses how statistical changepoint detection is used to identify when our complex system's performance characteristics have significantly changed. We will discuss the challenges of deploying this, such as dealing with "expected" changepoints like market open/close and downtime, relaying the change information to engineers in an effective manner, and establishing a feedback loop.

Joseph Cirella is a Software Engineer at Bloomberg, where he works within the Ticker Plant SRE Capacity & Performance team.

Shanthini Velan leads the Ticker Plant SRE Capacity & Performance team at Bloomberg.

BibTeX
@conference {272783,
author = {Joseph Cirella and Shanthini Velan},
title = {Stopping Performance Regression via Changepoint Detection},
year = {2025},
address = {Santa Clara, CA},
publisher = {USENIX Association},
month = mar
}