Releases: clowdhaus/cookiecluster
Releases · clowdhaus/cookiecluster
v0.5.0
0.5.0 - 2024-10-02
Bug Fixes
- Update test snapshots for latest instance types that were added
Dependencies
- Update dependencies to latest
Features
- Add support for EKS 1.31 and make it the default version
- Update list of available ec2 instance types
Miscellaneous Tasks
- Update version prior to
v0.5.0
release
v0.4.0
0.4.0 - 2024-09-11
Dependencies
- Update dependencies to the latest
Features
- Add support for
c7i-flex
,g6e
,hpc6a.48xlarge
,hpc6id.32xlarge
,hpc7a
, andp5e.48xlarge
instance types - Update EKS module version to latest, enable Karpenter v1 permissions
Miscellaneous Tasks
- Update version prior to
v0.4.0
release - Update test snapshots
v0.3.1
v0.3.0
v0.2.1
v0.2.0
0.2.0 - 2024-07-05
Bug Fixes
- Clarify Neuron device plugin does not have a Helm chart; Daemonset will need to be modified directly
- Remove default instance type selection, requiring users to de-select
Features
- Add notes to the generated output to guide users to additional steps needed for NVIDIA, Neuron, and EFA
- Add support for ML capacity block reservation with EKS managed node group
Miscellaneous Tasks
- Move VPC and subnet data sources to bottom of file to optimize the "above the fold"
- Update readme with explanation and demo
- Rename crate directory
cc
->cookiecluster
; updateREADME
v0.1.0
0.1.0 - 2024-07-04
Bug Fixes
- Correct spelling of
instance_types
and revert to private attribute - Correct choice filtering
Features
- Render instance storage and increased root volume
- Ensure only one instance is provided when using EFA or an EC2 capacity reservation
- Add support for integrating with existing network resources
- Add support for rendering Karpenter resources
- Render on-demand capacity reservation details
- Render NVIDIA and Neuron node groups
- Render instance types
- Feat: Render add-ons correctly w/ service account role ARN and
- Pass inputs to template and start rendering
- Add cluster version and add-ons
- Collect inputs and filter downstream selections based on each input step value
- Collect instance data from multiple regions and add non-GA instance types
- Add workload type selection; refactor input collection
- Add initial dialoguer using instance types
- Add xtask for collecting EC2 instance types to use in auto-completion
- Add initial template embedding and rendering
Miscellaneous Tasks
- Add CD release workflow
- Split up compute instance snapshot tests
- Add snapshot tests for rendered output
- Remove unsupported instance types
- Add tests
- Refactor inputs into sub-module
- Remove unused inputs and add description to help output
- Collapse the NVIDIA and Neuron templates into one accelerated MNG template
- Re-order inputs
- Split out NVIDIA and Neuron node groups to separate template files
- Remove workload type
Refactor
- Separate user input collection from logic in order to test logic
- Convert instance type into struct with attributes used for filtering