4–7 Nov 2019

The observability graph: Knowledge graphs for automated infrastructure observability (sponsored by Datadog)

Homin Lee (Datadog)
13:2514:05 Wednesday, 6 November 2019
Location: M1
Tags: wl
Average rating: **...
(2.00, 3 ratings)

Who is this presentation for?

  • Developers, SREs, architects, ML engineers, and data scientists

Level

Beginner

Description

The complexity and dynamism of distributed systems reflect both the teams that develop them and the products they support. Your observability tools need to be dynamic enough to keep up with your changing infrastructure and flexible enough to take in observability data of all types, yet understand that data well enough to show only what you care about.

Knowledge graphs have become an increasingly popular method for storing data, as they can explicitly but also flexibly encode the relationships between entities. Using both domain knowledge and user interaction data, Homin Lee walks you through how to train models to encode the vast amounts of data produced by observability tools into a knowledge graph, discover how they are interlinked, and imbue them with meaning.

You’ll see the details of how to construct these knowledge graphs, dubbed observability graphs, to reflect the innate structure (but also the uncertainty) in today’s dynamic infrastructure. Homin outlines how the knowledge graphs can be used to power automated alerting, alert clustering, and automated root-cause analysis.

This session is sponsored by Datadog.

Prerequisite knowledge

  • A basic understanding of observability tools

What you'll learn

  • Discover how to use knowledge graphs to improve your infrastructure observability
Photo of Homin Lee

Homin Lee

Datadog

Homin Lee is a data scientist at Datadog, where he writes algorithms that process trillions of data points a day. Previously, Homin built large-scale machine learning systems at several startups. Homin has a PhD from Columbia University in computational learning theory and was a Computing Innovation Fellow at the University of Texas at Austin.

Contact us

confreg@oreilly.com

For conference registration information and customer service

partners@oreilly.com

For more information on community discounts and trade opportunities with O’Reilly conferences

velocity@oreilly.com

For information on exhibiting or sponsoring a conference

pr@oreilly.com

For media/analyst press inquires