Skip to content

Enable OTel Container Insights in Windows EKS integ test harness - #343

Merged
sky333999 merged 1 commit into
mainfrom
fix/windows-eks-otel-container-insights
Jul 27, 2026
Merged

Enable OTel Container Insights in Windows EKS integ test harness#343
sky333999 merged 1 commit into
mainfrom
fix/windows-eks-otel-container-insights

Conversation

@sky333999

@sky333999 sky333999 commented Jul 23, 2026

Copy link
Copy Markdown
Contributor

Description

Fixes the failing EKS Win2019 / EKS Win2022 integration-test jobs.

Root cause

The shared windowslinux-tagged validation test (validations/eks/resources_generated_test.go) expects the OTel Container Insights resources to be present:

  • kube-state-metrics (Deployment, Service, ServiceAccount, ClusterRole, ClusterRoleBinding)
  • node-exporter (DaemonSet, Service, ServiceAccount, Role, RoleBinding)
  • cloudwatch-agent-cluster-scraper (operator-created Deployment + -monitoring Service)

All of these are gated in the chart behind otelContainerInsights.enabled, which defaults to false. The Linux EKS harness (terraform/eks/main.tf) sets otelContainerInsights.enabled=true, so it passes. The Windows harness (terraform/eks/windows/main.tf) never set it, so those resources were absent and TestResourcesGenerated failed on:

  • services: expected 13, got 10
  • deployments: expected 3, got 1
  • daemonSets: expected 8, got 7
  • 6 OTel CI RBAC Should be true checks

Change

  • Set otelContainerInsights.enabled=true in the Windows helm_release, matching the Linux harness.
  • Add rollout waits for the newly-enabled workloads (node-exporter DaemonSet, kube-state-metrics Deployment, cloudwatch-agent-cluster-scraper Deployment) before the validator runs.

Verification

helm template with --set otelContainerInsights.enabled=true (vs default) renders exactly the missing objects, accounting 1:1 for every failing assertion (+3 services, +2 deployments, +1 daemonset, +6 RBAC objects). Single Windows dir is parameterized by var.windows_os_version, so this covers both Win2019 and Win2022 jobs.

The windowslinux validation test expects OTel Container Insights resources
(kube-state-metrics, node-exporter, cloudwatch-agent-cluster-scraper), but the
Windows EKS Terraform harness never set otelContainerInsights.enabled (the Linux
harness does). This caused TestResourcesGenerated to fail on service/deployment/
daemonset counts and the OTel CI RBAC checks. Enable the flag and wait for the
newly-rendered workloads to roll out before validating.
@sky333999
sky333999 merged commit 333e805 into main Jul 27, 2026
40 checks passed
@sky333999
sky333999 deleted the fix/windows-eks-otel-container-insights branch July 27, 2026 19:21
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants