DevConf.cz 2022 has ended
Back To Schedule
Saturday, January 29 • 3:30pm - 3:55pm
Data Driven Resource Tiers for OpenShift Apps

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Ever wondered how much CPU or memory your application pod would need? Often, we find ourselves tempted to choose arbitrary and conventional numbers like 4 cores and 8 gigs of memory. In this talk, we are going to look at data driven resource tier recommendations for a common OpenShift application for data science users, Jupyterhub.

Jupyterhub provides an environment for running data science code in the cloud. Each user chooses from the configuration tiers arbitrarily set by the cluster administrator and the application then spawns their pods based on the selected tier. We propose an approach that first collects Telemetry data from user pods on the cluster. Then, it analyzes their usage patterns to recommend tiers that the administrator can use to make an informed decision before setting tiers for user pods.

Join us to learn how you can use telemetry data from your cluster to optimize it’s resource usage. The talk will introduce you to an application of AIOps for your cloud environment and show how data can be used to drive decisions.

Session chairs: Petr Muller and Viktor Malik

avatar for Shrey Anand

Shrey Anand

Data Scientist, Red Hat, Inc.
Shrey Anand is a data scientist with the AI Center of Excellence team at Red Hat. He's interested in defining and solving problems with system configuration data.

Saturday January 29, 2022 3:30pm - 3:55pm CET
Session Room 1