Optimize NVIDIA GPU operator jobs

* Move arm64 tests to the respective version files
* Skip gather steps on success
* Mark job as successful even if a best effort step failed
* Use smaller instance types
* Run arm64 jobs on amd64 test farms
* Remove unused gather steps

Author: empovit

ci-operator/config/rh-ecosystem-edge/nvidia-ci/.config.prowgen

@@ -14,10 +14,6 @@ slack_reporter:
       {{end}}
     {{end}}
   job_names:
-  - nvidia-gpu-operator-e2e-24-3-x
-  - nvidia-gpu-operator-e2e-24-6-x
-  - nvidia-gpu-operator-e2e-24-9-x
-  - nvidia-gpu-operator-e2e-24-12-x
   - nvidia-gpu-operator-e2e-25-3-x
   - nvidia-gpu-operator-e2e-25-10-x
   - nvidia-gpu-operator-e2e-master
@@ -26,3 +22,4 @@ slack_reporter:
   - nvidia-network-operator-shared-device-eth-rdma
   - mirror-vcsref-image
   - nvidia-dra-gpu-e2e
+  - nvidia-gpu-operator-e2e-arm64

ci-operator/config/rh-ecosystem-edge/nvidia-ci/rh-ecosystem-edge-nvidia-ci-main__4.14-stable.yaml

@@ -7,6 +7,14 @@ images:
 - dockerfile_path: Containerfile
   to: nvidia-ci
 releases:
+  arm64-latest:
+    prerelease:
+      architecture: arm64
+      product: ocp
+      version_bounds:
+        lower: 4.14.0-0
+        stream: 4-stable
+        upper: 4.15.0-0
   latest:
     prerelease:
       product: ocp
@@ -46,6 +54,19 @@ tests:
       BASE_DOMAIN: edge-sro.rhecoeng.com
       NVIDIAGPU_SUBSCRIPTION_CHANNEL: v25.3
     workflow: nvidia-gpu-operator-e2e-aws
+- always_run: false
+  as: nvidia-gpu-operator-e2e-arm64
+  steps:
+    cluster_profile: aws-edge-infra
+    dependencies:
+      OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE: release:arm64-latest
+    env:
+      AWS_REGION_OVERWRITE: us-east-1
+      BASE_DOMAIN: edge-sro.rhecoeng.com
+      COMPUTE_NODE_TYPE: g5g.2xlarge
+      CONTROL_PLANE_INSTANCE_TYPE: m6g.xlarge
+      OCP_ARCH: arm64
+    workflow: nvidia-gpu-operator-e2e-aws
 zz_generated_metadata:
   branch: main
   org: rh-ecosystem-edge

ci-operator/config/rh-ecosystem-edge/nvidia-ci/rh-ecosystem-edge-nvidia-ci-main__4.15-stable.yaml

@@ -7,6 +7,14 @@ images:
 - dockerfile_path: Containerfile
   to: nvidia-ci
 releases:
+  arm64-latest:
+    prerelease:
+      architecture: arm64
+      product: ocp
+      version_bounds:
+        lower: 4.15.0-0
+        stream: 4-stable
+        upper: 4.16.0-0
   latest:
     prerelease:
       product: ocp
@@ -46,6 +54,19 @@ tests:
       BASE_DOMAIN: edge-sro.rhecoeng.com
       NVIDIAGPU_SUBSCRIPTION_CHANNEL: v25.3
     workflow: nvidia-gpu-operator-e2e-aws
+- always_run: false
+  as: nvidia-gpu-operator-e2e-arm64
+  steps:
+    cluster_profile: aws-edge-infra
+    dependencies:
+      OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE: release:arm64-latest
+    env:
+      AWS_REGION_OVERWRITE: us-east-1
+      BASE_DOMAIN: edge-sro.rhecoeng.com
+      COMPUTE_NODE_TYPE: g5g.2xlarge
+      CONTROL_PLANE_INSTANCE_TYPE: m6g.xlarge
+      OCP_ARCH: arm64
+    workflow: nvidia-gpu-operator-e2e-aws
 zz_generated_metadata:
   branch: main
   org: rh-ecosystem-edge

ci-operator/config/rh-ecosystem-edge/nvidia-ci/rh-ecosystem-edge-nvidia-ci-main__4.16-stable.yaml

@@ -7,6 +7,14 @@ images:
 - dockerfile_path: Containerfile
   to: nvidia-ci
 releases:
+  arm64-latest:
+    prerelease:
+      architecture: arm64
+      product: ocp
+      version_bounds:
+        lower: 4.16.0-0
+        stream: 4-stable
+        upper: 4.17.0-0
   latest:
     prerelease:
       product: ocp
@@ -46,6 +54,19 @@ tests:
       BASE_DOMAIN: edge-sro.rhecoeng.com
       NVIDIAGPU_SUBSCRIPTION_CHANNEL: v25.3
     workflow: nvidia-gpu-operator-e2e-aws
+- always_run: false
+  as: nvidia-gpu-operator-e2e-arm64
+  steps:
+    cluster_profile: aws-edge-infra
+    dependencies:
+      OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE: release:arm64-latest
+    env:
+      AWS_REGION_OVERWRITE: us-east-1
+      BASE_DOMAIN: edge-sro.rhecoeng.com
+      COMPUTE_NODE_TYPE: g5g.2xlarge
+      CONTROL_PLANE_INSTANCE_TYPE: m6g.xlarge
+      OCP_ARCH: arm64
+    workflow: nvidia-gpu-operator-e2e-aws
 zz_generated_metadata:
   branch: main
   org: rh-ecosystem-edge

ci-operator/config/rh-ecosystem-edge/nvidia-ci/rh-ecosystem-edge-nvidia-ci-main__4.17-stable.yaml

@@ -7,6 +7,14 @@ images:
 - dockerfile_path: Containerfile
   to: nvidia-ci
 releases:
+  arm64-latest:
+    prerelease:
+      architecture: arm64
+      product: ocp
+      version_bounds:
+        lower: 4.17.0-0
+        stream: 4-stable
+        upper: 4.18.0-0
   latest:
     prerelease:
       product: ocp
@@ -46,6 +54,19 @@ tests:
       BASE_DOMAIN: edge-sro.rhecoeng.com
       NVIDIAGPU_SUBSCRIPTION_CHANNEL: v25.3
     workflow: nvidia-gpu-operator-e2e-aws
+- always_run: false
+  as: nvidia-gpu-operator-e2e-arm64
+  steps:
+    cluster_profile: aws-edge-infra
+    dependencies:
+      OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE: release:arm64-latest
+    env:
+      AWS_REGION_OVERWRITE: us-east-1
+      BASE_DOMAIN: edge-sro.rhecoeng.com
+      COMPUTE_NODE_TYPE: g5g.2xlarge
+      CONTROL_PLANE_INSTANCE_TYPE: m6g.xlarge
+      OCP_ARCH: arm64
+    workflow: nvidia-gpu-operator-e2e-aws
 zz_generated_metadata:
   branch: main
   org: rh-ecosystem-edge

ci-operator/config/rh-ecosystem-edge/nvidia-ci/rh-ecosystem-edge-nvidia-ci-main__4.18-stable.yaml

@@ -7,6 +7,14 @@ images:
 - dockerfile_path: Containerfile
   to: nvidia-ci
 releases:
+  arm64-latest:
+    prerelease:
+      architecture: arm64
+      product: ocp
+      version_bounds:
+        lower: 4.18.0-0
+        stream: 4-stable
+        upper: 4.19.0-0
   latest:
     prerelease:
       product: ocp
@@ -46,6 +54,19 @@ tests:
       BASE_DOMAIN: edge-sro.rhecoeng.com
       NVIDIAGPU_SUBSCRIPTION_CHANNEL: v25.3
     workflow: nvidia-gpu-operator-e2e-aws
+- always_run: false
+  as: nvidia-gpu-operator-e2e-arm64
+  steps:
+    cluster_profile: aws-edge-infra
+    dependencies:
+      OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE: release:arm64-latest
+    env:
+      AWS_REGION_OVERWRITE: us-east-1
+      BASE_DOMAIN: edge-sro.rhecoeng.com
+      COMPUTE_NODE_TYPE: g5g.2xlarge
+      CONTROL_PLANE_INSTANCE_TYPE: m6g.xlarge
+      OCP_ARCH: arm64
+    workflow: nvidia-gpu-operator-e2e-aws
 zz_generated_metadata:
   branch: main
   org: rh-ecosystem-edge

ci-operator/config/rh-ecosystem-edge/nvidia-ci/rh-ecosystem-edge-nvidia-ci-main__4.19-stable.yaml

@@ -7,6 +7,14 @@ images:
 - dockerfile_path: Containerfile
   to: nvidia-ci
 releases:
+  arm64-latest:
+    prerelease:
+      architecture: arm64
+      product: ocp
+      version_bounds:
+        lower: 4.19.0-0
+        stream: 4-stable
+        upper: 4.20.0-0
   latest:
     prerelease:
       product: ocp
@@ -46,6 +54,19 @@ tests:
       BASE_DOMAIN: edge-sro.rhecoeng.com
       NVIDIAGPU_SUBSCRIPTION_CHANNEL: v25.3
     workflow: nvidia-gpu-operator-e2e-aws
+- always_run: false
+  as: nvidia-gpu-operator-e2e-arm64
+  steps:
+    cluster_profile: aws-edge-infra
+    dependencies:
+      OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE: release:arm64-latest
+    env:
+      AWS_REGION_OVERWRITE: us-east-1
+      BASE_DOMAIN: edge-sro.rhecoeng.com
+      COMPUTE_NODE_TYPE: g5g.2xlarge
+      CONTROL_PLANE_INSTANCE_TYPE: m6g.xlarge
+      OCP_ARCH: arm64
+    workflow: nvidia-gpu-operator-e2e-aws
 zz_generated_metadata:
   branch: main
   org: rh-ecosystem-edge

ci-operator/config/rh-ecosystem-edge/nvidia-ci/rh-ecosystem-edge-nvidia-ci-main__4.20-stable.yaml

@@ -7,6 +7,14 @@ images:
 - dockerfile_path: Containerfile
   to: nvidia-ci
 releases:
+  arm64-latest:
+    prerelease:
+      architecture: arm64
+      product: ocp
+      version_bounds:
+        lower: 4.20.0-0
+        stream: 4-stable
+        upper: 4.21.0-0
   latest:
     prerelease:
       product: ocp
@@ -46,6 +54,19 @@ tests:
       BASE_DOMAIN: edge-sro.rhecoeng.com
       NVIDIAGPU_SUBSCRIPTION_CHANNEL: v25.3
     workflow: nvidia-gpu-operator-e2e-aws
+- always_run: false
+  as: nvidia-gpu-operator-e2e-arm64
+  steps:
+    cluster_profile: aws-edge-infra
+    dependencies:
+      OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE: release:arm64-latest
+    env:
+      AWS_REGION_OVERWRITE: us-east-1
+      BASE_DOMAIN: edge-sro.rhecoeng.com
+      COMPUTE_NODE_TYPE: g5g.2xlarge
+      CONTROL_PLANE_INSTANCE_TYPE: m6g.xlarge
+      OCP_ARCH: arm64
+    workflow: nvidia-gpu-operator-e2e-aws
 zz_generated_metadata:
   branch: main
   org: rh-ecosystem-edge

ci-operator/config/rh-ecosystem-edge/nvidia-ci/rh-ecosystem-edge-nvidia-ci-main__4.21-stable.yaml

@@ -7,6 +7,14 @@ images:
 - dockerfile_path: Containerfile
   to: nvidia-ci
 releases:
+  arm64-latest:
+    prerelease:
+      architecture: arm64
+      product: ocp
+      version_bounds:
+        lower: 4.21.0-0
+        stream: 4-stable
+        upper: 4.22.0-0
   latest:
     prerelease:
       product: ocp
@@ -46,6 +54,19 @@ tests:
       BASE_DOMAIN: edge-sro.rhecoeng.com
       NVIDIAGPU_SUBSCRIPTION_CHANNEL: v25.3
     workflow: nvidia-gpu-operator-e2e-aws
+- always_run: false
+  as: nvidia-gpu-operator-e2e-arm64
+  steps:
+    cluster_profile: aws-edge-infra
+    dependencies:
+      OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE: release:arm64-latest
+    env:
+      AWS_REGION_OVERWRITE: us-east-1
+      BASE_DOMAIN: edge-sro.rhecoeng.com
+      COMPUTE_NODE_TYPE: g5g.2xlarge
+      CONTROL_PLANE_INSTANCE_TYPE: m6g.xlarge
+      OCP_ARCH: arm64
+    workflow: nvidia-gpu-operator-e2e-aws
 zz_generated_metadata:
   branch: main
   org: rh-ecosystem-edge

ci-operator/config/rh-ecosystem-edge/nvidia-ci/rh-ecosystem-edge-nvidia-ci-main__4.22-stable.yaml

@@ -7,6 +7,14 @@ images:
 - dockerfile_path: Containerfile
   to: nvidia-ci
 releases:
+  arm64-latest:
+    prerelease:
+      architecture: arm64
+      product: ocp
+      version_bounds:
+        lower: 4.22.0-0
+        stream: 4-stable
+        upper: 4.23.0-0
   latest:
     prerelease:
       product: ocp
@@ -51,6 +59,21 @@ tests:
       NVIDIAGPU_GPU_FALLBACK_CATALOGSOURCE_INDEX_IMAGE: registry.redhat.io/redhat/certified-operator-index:v4.21
       NVIDIAGPU_SUBSCRIPTION_CHANNEL: v25.3
     workflow: nvidia-gpu-operator-e2e-aws
+- always_run: false
+  as: nvidia-gpu-operator-e2e-arm64
+  steps:
+    cluster_profile: aws-edge-infra
+    dependencies:
+      OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE: release:arm64-latest
+    env:
+      AWS_REGION_OVERWRITE: us-east-1
+      BASE_DOMAIN: edge-sro.rhecoeng.com
+      COMPUTE_NODE_TYPE: g5g.2xlarge
+      CONTROL_PLANE_INSTANCE_TYPE: m6g.xlarge
+      NFD_FALLBACK_CATALOGSOURCE_INDEX_IMAGE: registry.redhat.io/redhat/redhat-operator-index:v4.21
+      NVIDIAGPU_GPU_FALLBACK_CATALOGSOURCE_INDEX_IMAGE: registry.redhat.io/redhat/certified-operator-index:v4.21
+      OCP_ARCH: arm64
+    workflow: nvidia-gpu-operator-e2e-aws
 zz_generated_metadata:
   branch: main
   org: rh-ecosystem-edge