Loading…
Attending this event?
Amsterdam, The Netherlands
March 30–April 2, 2020

Learn More and Register to Attend This Event
Wednesday, April 1 • 11:50 - 12:25
Operating Kubernetes at Pinterest: Past, Present and Future - Rodrigo Menezes & Roberto Rodriguez Alcala, Pinterest

Sign up or log in to save this to your schedule and see who's attending!

Over the last 2 years, Pinterest has been working on migrating systems to Kubernetes.
Currently we operate multiple multi tenant clusters that run ML workloads on GPUs and CPUs as well as highly available online serving systems. Operating Kubernetes at this scale has presented some challenges, like upgrades, safe deployments, autoscaling and more.

By leveraging Kops and combining it with other tools, we've been able to successfully overcome some of the challenges and we can now do things like
set up a cluster in hours, easily reshape our infrastructure or roll out updates seamlessly.

Along the way we added extra features like ephemeral nodes using spot instances, huge pages, GPUs and more; without any significant impact on the operation load.

This session will go through the design decisions, the reasons behind, details about our implementation and improvements that we're working on.

Speakers
avatar for Rodrigo Menezes

Rodrigo Menezes

Site Reliability Engineer, Pinterest
Rodrigo is a member of the Core Site Reliability Engineering team at Pinterest. While at Pinterest, his main focus has been Kubernetes and creating applications to support running Pinterest's stateless prod infrastructure in containers. Outside of work, Rodrigo loves rock climbing... Read More →
avatar for Roberto Rodriguez Alcala

Roberto Rodriguez Alcala

Site Reliability Engineer, Pinterest
Roberto is a Site Reliability Engineer embedded in the Cloud Management Platform team at Pinterest. His main focus has been on Kubernetes, developing applications and putting building blocks together to allow existing workloads to migrate from existing infra to our Kubernetes clusters... Read More →


Wednesday April 1, 2020 11:50 - 12:25
TBA
Feedback form isn't open yet.

Attendees (33)