Deploying JupyterHub with Kubernetes

on YouTube ↗27m

" summary

Shows how a production JupyterHub is built on Kubernetes with concrete choices: GKE + Helm chart for quick deploy, alternatives on Azure and DIY on AWS, and a live demo cluster running on Google Cloud. Gives costs and scale: €80/month for three small nodes example, cites Data 8: 1000+ users and a case of ~100k users using multiple Hubs and Postgres for state. Explains hub proxy single-user servers, resource isolation with containers, and autoscaling tradeoffs.

AI-written summary · about

" chapters (9)

1. 0:00JupyterHub and cloud deployment
2. 4:13What JupyterHub is and challenges
3. 8:55JupyterHub production challenges
4. 10:19Containers for deployment issues
5. 13:29Kubernetes for JupyterHub
6. 16:36Deploying JupyterHub on cloud
7. 20:50Scaling JupyterHub for users
8. 22:57Summary
9. 23:19Q&A

" related talks

• 37mJupyterHub from the Ground Up with KubernetesCamilla Montonen · PyData London 2018
• 39mHow We Hosted 200+ Data Scientists in a K8S ClusterAleksey Kuznetsov, Mikhail Lepeshkin · Saint HighLoad++ 2022
• 35mLarge Scale Teaching Infrastructure with KubernetesYuvi Panda · KubeCon + CloudNativeCon North America 2017
• 26mTerraforming Jupyter and Dask: How to Get Your Own Cloud Data Science Platform on the CheapChris Ostrouchov · HashiTalks 2021
• 31mBuilding an interactive training environment using JupyterHubGraham Dumpleton · PyCon AU 2019
• 19mOptimizing Resilience and Availability by Migrating from JupyterHub to the...Alexander Perlman, David Hoover · SREcon24 Americas
• 20mKeynote: Kubernetes at GitHubJesse Newland · KubeCon + CloudNativeCon North America 2017
• 32mDemocratizing Distributed Computing with Dask and JupyterHubMatthew Rocklin · PyCon 2018