You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
My (hazy) understanding of what this is is a way to deploy a cluster in the cloud on-demand and submit work to it without having a client process controlling it at all times. This seems like it could be useful as a way to provision a cluster on which to run an in-memory rechunk without having to worry about turning off the cluster again.
The text was updated successfully, but these errors were encountered:
Note that you can do this without Docker, as it works with Coiled's Package Sync. The spin-up time is said to be ~30-60s.
There are now two interesting possibilities for running Cubed on Coiled without Dask* which makes Coiled much more similar to Modal for our deployment purposes.
*The other being to run with the single-machine executor on a single fat instance that you requisition via a Coiled notebook.
One of Matt Rocklin's blog posts mentions Prefect's "serverless work pool"
https://docs.prefect.io/v3/deploy/infrastructure-examples/serverless
My (hazy) understanding of what this is is a way to deploy a cluster in the cloud on-demand and submit work to it without having a client process controlling it at all times. This seems like it could be useful as a way to provision a cluster on which to run an in-memory rechunk without having to worry about turning off the cluster again.
The text was updated successfully, but these errors were encountered: