-
Notifications
You must be signed in to change notification settings - Fork 182
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We鈥檒l occasionally send you account related emails.
Already on GitHub? Sign in to your account
Enhance pull speed for Large ML container Images with Bottlerocket #559
Comments
I would like to work on this |
@lindarr915 Please let us know if you need any guidance on the repo. This will be a great addition. You can also write a blog doc under the resources section to discuss the performance improvements with Bottlerocket. |
Hi @vara-bonthu Darren works with me in a same team. One thing we have discussed, do you want to make this solution as a parallel pattern as others? or make it kind of a "shared pattern" (eg. a TF module or a separate stack), as some of the existing patterns may all need it. |
Community Note
What is the outcome that you are trying to reach?
Currently the inference container images run into multiple GBs in size. This negatively impacts the start up time for Ray Pods. We should look for ways to reduce the startup time of the Ray head and worker pods.
Describe the solution you would like
Validate if pre-fetching of images can be used leveraging EKS bottle rocket data volume in the inference blueprints to achieve this.
Describe alternatives you have considered
Additional context
https://aws.amazon.com/blogs/containers/reduce-container-startup-time-on-amazon-eks-with-bottlerocket-data-volume/
The text was updated successfully, but these errors were encountered: