feat: add NODE_TYPE, EKS_VERSION, CLUSTER_TAGS, USE_SPOT options to AWS EKS setup#349
Conversation
michaelraney
requested review from
WentingWu666666,
alaye-ms,
hossain-rayhan and
xgerman
as code owners
michaelraney
force-pushed
the
feat/aws-observability-addon-migration
branch
5 times, most recently
from
3c33783 to
a2888a5
Compare
michaelraney
changed the title
|
Hi @michaelraney, Thanks for the contribution! Great to see someone improving the AWS EKS experience. Quick thought on scope: we try to keep Suggestion: Let's keep the straightforward options in this PR: NODE_TYPE, EKS_VERSION, CLUSTER_TAG, USE_SPOT. For the advanced stuff (VPC endpoints, Container Insights, event filtering), would you mind splitting those into a separate PR under |
…WS EKS setup Add cost-oriented defaults and controls to the AWS EKS playground: - NODE_TYPE default bumped to m7g.large (Graviton/ARM) with --node-type override - K8S_VERSION pinned to 1.35 with --eks-version override - --spot opt-in for Spot-backed managed nodes (dev/test only) - --tags (and CLUSTER_TAGS env var) for AWS cost allocation tracking Scope is intentionally limited per reviewer feedback on PR documentdb#349; advanced observability and VPC-endpoint work is delivered in a separate contrib PR. Signed-off-by: michaelraney <raneymike83@gmail.com> Made-with: Cursor
michaelraney
force-pushed
the
feat/aws-observability-addon-migration
branch
from
a2888a5 to
1d466c8
Compare
michaelraney
changed the title
|
Thanks @hossain-rayhan — split done as suggested. This PR is now narrowed to just the four straightforward options you called out: The advanced CloudWatch observability and cost-optimization work (S3 Gateway VPC endpoint, EKS control-plane logging + retention, Amazon CloudWatch Observability add-on / Container Insights, CloudWatch-aware teardown and diagnostics, 2-AZ deployment, mongosh prereq, CloudFormation event diagnostics) moved to a new self-contained contrib variant: #352 ( |
| NODE_TYPE="$2" | ||
| shift 2 | ||
| ;; | ||
| --eks-version|--k8s-version) |
There was a problem hiding this comment.
Are the eks-version and k8s-version same term internally for AWS customers?
2 participants
Summary
Adds narrowly-scoped cost-optimization options to the AWS EKS playground, per reviewer feedback to keep
documentdb-playground/aws-setup/simple and DocumentDB-focused:--node-type/NODE_TYPEenv var — EC2 instance type (default:m7g.large, Graviton/ARM).--eks-version/K8S_VERSIONenv var — Kubernetes/EKS version (default:1.35).--spot/USE_SPOTenv var — opt-in Spot-backed managed nodes for dev/test cost reduction (~70%).--tags/CLUSTER_TAGSenv var — AWS cost allocation tags for Cost Explorer tracking.Why
The EKS setup should be affordable by default and easy to attribute in Cost Explorer. These four knobs cover the straightforward options without introducing AWS-specific infrastructure the core team would need to maintain.
Scope split
The advanced work originally in this PR (S3 Gateway VPC endpoint, Amazon CloudWatch Observability add-on / Container Insights, EKS control-plane logging, log-group retention, CloudWatch-aware teardown and diagnostics) has been moved to a separate follow-up PR under
documentdb-playground/contrib/telemetry-and-cost-optimized-eks/so the core scripts stay simple, while users who want the full telemetry+cost-optimized variant can opt in.Test plan
./documentdb-playground/aws-setup/scripts/create-cluster.shwith defaults — cluster created on m7g.large with default tags applied../documentdb-playground/aws-setup/scripts/create-cluster.sh --spot— Spot-backed nodes provisioned, warning banner shown../documentdb-playground/aws-setup/scripts/create-cluster.sh --tags "project=myproj,team=platform"— custom tags present on AWS resources../documentdb-playground/aws-setup/scripts/create-cluster.sh --eks-version 1.34 --node-type m5.large— overrides respected../documentdb-playground/aws-setup/scripts/delete-cluster.sh— teardown completes cleanly (delete-cluster.sh is unchanged in this PR).