gke shutdown steps #2565
Comments
|
Thank you for opening your first issue in this project! Engagement like this is essential for open source projects! 🤗 |
|
Writing down a record of
|
|
And snapshot of resource metrics for the last 30 days: https://snapshots.raintank.io/dashboard/snapshot/eZBW6PGLeGULWghRSDzqGx4hmTqII7ie |
|
gke.mybinder.org is shutdown, all images are deleted, and prod is currently running on a single e2-highmem-2 node. |
|
@minrk can we close this one? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Tracking more fine-grained tasks in this repo to shutdown GKE:
part of jupyterhub/team-compass#642
staging fits comfortably on a single e2-standard-2 (~$50-70/month) with no management cost because the cluster is zonal.
Prod is a regional cluster, which costs $75/month. I think we can fit prod on e2-standard-2 or e2-standard-4 core nodes. Prometheus is the big mystery here - prometheus on current prod uses a lot of resources, because it needs to handle tons of launches. That won't be the case anymore, but it's unclear to me how much it can be reduced, because of the resources required by traffic to the federation-redirect entrypoint.
Two e2-standard-4 costs ~$220/month, for a total around $300/month for prod. If we can fit on one e2-standard-4 or two e2-standard-2, it would be more like $200.
The text was updated successfully, but these errors were encountered: