feat(ai-platform): SAIA service exposure to external requests#86
Merged
Conversation
- cluster-config.yaml: rewrite SAIA exposure as 3 NodePort-free modes, drop redundant nginx image entry; add byoTargetGroup config block - eks_cluster_with_stack.sh: read BYO_TG_*; add validate_byo_target_group_config, apply_byo_target_group_binding, patch_saia_service_disable_nodeport; update patch_saia_public_service_workaround for NodePort-free mode - k0s-cluster-config.yaml: switch SAIA exposure to LoadBalancer + MetalLB; add metallb config block; revert object storage to type=minio with AWS S3 endpoint (the only working path on k0s — type=aws is silently swapped to in-cluster MinIO by the install script) - k0s_cluster_with_stack.sh: add install_metallb function (chart pin 0.14.8, L2 / BGP advertisements); patch_k0s_saia_service_disable_nodeport; fix describe_pod node-count whitespace bug - artifacts.yaml: minor diff (will be overwritten by upcoming merge) Pre-merge of origin/ai-tier-v2-k0s; will be subsumed by the merge commit. Made-with: Cursor
Brings in: - 4146385 refactor: remove in-cluster MinIO install (k0s installer now requires customer-managed object storage; aligns with the bug we hit where type=aws was silently swapped to in-cluster MinIO) - 86cf822 fix: removal of aws specific usages - 880f68b feature: including saia deployments helm configs - 3d1104d feat: added initContainer for saia-vector-db-setup posthook - 922cb4f fix: add safety gate to prevent install_k0s_cluster from wiping a live cluster - 0ccde9f refactor: remove ecr credential refresher - d74d9c5 fix: github copilot review comments - plus go.mod/go.sum CVE patches and minor cleanups Conflict resolutions: - tools/cluster_setup/artifacts.yaml: take theirs (generated CRD bundle from new operator) - tools/cluster_setup/cluster-config.yaml: block 1: keep our LIFECYCLE WORKFLOW comment, take theirs' placeholder values (useExisting: false, name: my-ai-cluster) block 2: keep ours (objectStore.type=aws — EKS supports AWS S3 end-to-end) - tools/cluster_setup/k0s-cluster-config.yaml: sshKeyPath: take theirs (~/.ssh/id_rsa placeholder) existingIPs: take theirs (10.0.0.x placeholder IPs) storage header: take theirs (new preflight checks comment) objectStore: keep ours (type=minio + AWS S3 endpoint — the only working path on k0s, per the post-merge testing) operator image: take theirs (splunk-ai-operator:latest placeholder) files paths: take theirs (with CHANGE THIS comments) - tools/cluster_setup/k0s_cluster_with_stack.sh: auto-merged cleanly (their MinIO-removal refactor + our install_metallb / disable-NodePort function landed in different sections, no manual fixup needed) Validation post-merge: - bash -n passes on both eks/k0s install scripts - yq parses all three cluster configs - zero conflict markers anywhere - key sections preserved on both sides: EKS: objectStore.type=aws, serviceTemplate.type=LoadBalancer, byoTargetGroup.enabled=false (new field) k0s: objectStore.type=minio (CRD workaround), metallb.install=true (our MetalLB block), install_metallb() function intact Made-with: Cursor
Contributor
There was a problem hiding this comment.
Pull request overview
Adds an environment-agnostic contract for exposing the SAIA public Service to external clients, wiring installer scripts + config templates to support NodePort-free external access (MetalLB on k0s; AWS Load Balancer Controller / TargetGroupBinding on EKS) without requiring operator changes.
Changes:
- Add guardrails against applying placeholder object-store credentials and improve preflight validation.
- Implement NodePort-free SAIA exposure flows: MetalLB install/patching for k0s; AWS LBC install + optional BYO target group binding for EKS.
- Update cluster config templates and installer robustness tweaks (macOS sed handling, EIP/NAT guidance, health check behavior).
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.
| File | Description |
|---|---|
| tools/cluster_setup/k0s_cluster_with_stack.sh | Adds placeholder credential detection, MetalLB install + SAIA Service patching flow, and adjusts health check logic. |
| tools/cluster_setup/k0s-cluster-config.yaml | Updates k0s template to default to LoadBalancer exposure + adds MetalLB configuration + placeholder creds. |
| tools/cluster_setup/eks_cluster_with_stack.sh | Adds SAIA serviceTemplate handling, AWS Load Balancer Controller install/IRSA + BYO TargetGroupBinding mode, and image wiring updates. |
| tools/cluster_setup/cluster-config.yaml | Updates EKS config template to document/configure NodePort-free SAIA exposure modes and related AWS settings. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
kupratyu-splunk
approved these changes
Apr 29, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
Adds a single, environment-agnostic contract for exposing the SAIA
public Service to external clients. The operator already renders a
public Service named
<aiPlatform.name>-saia-servicewhose endpointsare the in-cluster nginx pods (which terminate path-based v1/v2
routing). This change wires up the install scripts and cluster-config
template so the same Service can be reached three different ways
without any operator code change:
Related Issues
Type of Change
Changes Made
Testing Performed
make test)make lint)Test Environment
Test Steps
Documentation
Checklist
Breaking Changes
Impact:
Migration Path:
Screenshots/Recordings
Additional Notes
Reviewer Notes
Please pay special attention to:
Commit Message Convention: This PR follows Conventional Commits