Kafka at the New York Times and Datadog

/!\ DATE CHANGE: sorry for the late notice but we have to postpone this meetup. New date is Tuesday, March 12th. Hope you can all make it!

---------- TALKS ----------

---- How to rewind The New York Times homepage, and other things learned about running Kafka as a service ----

The Publishing Pipeline team at the New York Times stores news content permanently in an ordered, single-partition topic on Kafka. Backend teams use this topic as a platform for constructing data stores and services for various use cases. The team has spent the past two years evolving this architecture. In this talk, we will share the good, the bad, and the downright scary.

(Mike Kaminski + Stephen Dotz)

---- Thoughts on Kafka Capacity Planning ----

The cost of inefficiencies and risk of outages are magnified with scale. This talk will share a few perspectives on how capacity planning can be better understood in the context of Kafka with the goal of proactive rather than reactive measures.

(Jamie Alquiza)

---------- SPEAKERS ----------

Mike Kaminski is a Senior Software Engineer at The New York Times who has been working on the Publishing Pipeline team for 3 years and takes great pride in keeping the brokers ISR.

Stephen Dotz is Engineering Manager of the Publishing Pipeline team at The New York Times. In his past at NYT, he has worked on earlier iterations of publishing infrastructure, Apple News, and data analysis pipelines using BigQuery.

Jamie Alquiza is a Sr. Software Engineer working on high throughput datastores at Datadog.