We have deployed z2jh (0.11.1) on IBM’s Kubernetes services (15 large nodes), with about 20 active servers at any moment… but noticed, in general, spawning a JupterLab server takes 30sec and up to 60sec. Is this typical and are there ways to improve the performance of the spawner? Thanks for any help!
There are some suggestions on this page: Optimizations — Zero to JupyterHub with Kubernetes documentation
However I’d start by upgrading to the latest release, your Z2JH version is quite old.
For comparison how long would you expect your Kubernetes cluster to take to launch a typical non-Jupyter pod?
Thank you for your reply!
I deployed a pod with our JupyterLab image with kubectl and it took about 3-5 sec to launch… I would expect most pods to be below 5 seconds usually.
To clarify, the Jhub version is 1.3 and helm version 0.11.1, which is old, and we will try to update our YAML to use the latest.
Thank you for the link. We will try to walk through these and see if any will work for our situation.