Initial transition to applications from helm for applications products;

k8s-inference/helm.tf

@@ -3,7 +3,9 @@ module "gpu-operator" {
     nebius_mk8s_v1_node_group.gpu,
     nebius_mk8s_v1_node_group.cpu-only,
   ]
-  source = "../modules/gpu-operator"
+  source     = "../modules/gpu-operator"
+  parent_id  = var.parent_id
+  cluster_id = nebius_mk8s_v1_cluster.k8s-cluster.id
 }
 
 module "o11y" {

k8s-training/helm.tf

@@ -3,14 +3,18 @@ module "network-operator" {
     nebius_mk8s_v1_node_group.cpu-only,
     nebius_mk8s_v1_node_group.gpu,
   ]
-  source = "../modules/network-operator"
+  source     = "../modules/network-operator"
+  parent_id  = var.parent_id
+  cluster_id = nebius_mk8s_v1_cluster.k8s-cluster.id
 }
 
 module "gpu-operator" {
   depends_on = [
     module.network-operator
   ]
-  source = "../modules/gpu-operator"
+  source     = "../modules/gpu-operator"
+  parent_id  = var.parent_id
+  cluster_id = nebius_mk8s_v1_cluster.k8s-cluster.id
 }
 
 module "o11y" {
@@ -44,34 +48,34 @@ module "o11y" {
   }
   test_mode = var.test_mode
 }
-
-module "nccl-test" {
-  depends_on = [
-    module.gpu-operator,
-  ]
-
-  count           = var.test_mode ? 1 : 0
-  source          = "../modules/nccl-test"
-  number_of_hosts = nebius_mk8s_v1_node_group.gpu.fixed_node_count
-}
-
-module "kuberay" {
-  source = "../modules/kuberay"
-  count  = var.enable_kuberay ? 1 : 0
-
-  depends_on = [
-    nebius_mk8s_v1_node_group.cpu-only,
-    nebius_mk8s_v1_node_group.gpu,
-    module.network-operator,
-    module.gpu-operator,
-    module.csi-mounted-fs-path,
-  ]
-
-  gpu_platform     = var.gpu_nodes_platform
-  cpu_platform     = var.cpu_nodes_platform
-  min_gpu_replicas = var.kuberay_min_gpu_replicas
-  max_gpu_replicas = var.kuberay_max_gpu_replicas
-}
+#
+#module "nccl-test" {
+#  depends_on = [
+#    module.gpu-operator,
+#  ]
+#
+#  count           = var.test_mode ? 1 : 0
+#  source          = "../modules/nccl-test"
+#  number_of_hosts = nebius_mk8s_v1_node_group.gpu.fixed_node_count
+#}
+#
+#module "kuberay" {
+#  source = "../modules/kuberay"
+#  count  = var.enable_kuberay ? 1 : 0
+#
+#  depends_on = [
+#    nebius_mk8s_v1_node_group.cpu-only,
+#    nebius_mk8s_v1_node_group.gpu,
+#    module.network-operator,
+#    module.gpu-operator,
+#    module.csi-mounted-fs-path,
+#  ]
+#
+#  gpu_platform     = var.gpu_nodes_platform
+#  cpu_platform     = var.cpu_nodes_platform
+#  min_gpu_replicas = var.kuberay_min_gpu_replicas
+#  max_gpu_replicas = var.kuberay_max_gpu_replicas
+#}
 
 
 module "csi-mounted-fs-path" {

modules/gpu-operator/helm.tf

@@ -1,17 +1,19 @@
-resource "helm_release" "gpu-operator" {
-  name = "gpu-operator"
-  # FIXME set to production product
-  repository       = var.helm_repository
-  chart            = "gpu-operator"
-  namespace        = "gpu-operator"
-  create_namespace = true
-  version          = var.helm_version
-  atomic           = true
-  timeout          = 600
-
-  set {
-    name  = "driver.version"
-    value = var.driver_version
+resource "nebius_applications_v1alpha1_k8s_release" "gpu-operator" {
+  lifecycle {
+    ignore_changes = all
   }
+  cluster_id = var.cluster_id
+  parent_id  = var.parent_id
+
+  application_name = "gpu-operator"
+  namespace        = "gpu-operator"
+  product_slug     = "nebius/nvidia-gpu-operator"
 
+  #  set = {
+  #    "driver.version" : var.driver_version
+  #  }
+  values = <<EOT
+  driver:
+    version: ${var.driver_version}
+  EOT
 }

modules/gpu-operator/variables.tf

@@ -1,13 +1,17 @@
-variable "helm_repository" {
-  description = "GPU Operator Helm chart source repository."
+variable "cluster_id" {
+  description = "K8s cluster id."
   type        = string
-  default     = "oci://cr.eu-north1.nebius.cloud/marketplace/nebius/nvidia-gpu-operator/chart"
 }
 
-variable "helm_version" {
-  description = "Version of GPU Operator Helm chart."
+variable "parent_id" {
+  description = "Project id."
   type        = string
-  default     = "v24.6.2"
+}
+
+variable "product_slug" {
+  description = "Marketplace product slug."
+  type        = string
+  default     = "nebius/nvidia-gpu-operator"
 }
 
 variable "driver_version" {

modules/gpu-operator/versions.tf

@@ -0,0 +1,7 @@
+terraform {
+  required_providers {
+    nebius = {
+      source = "terraform-provider-nebius.storage.ai.nebius.cloud/nebius/nebius"
+    }
+  }
+}

modules/network-operator/helm.tf

@@ -1,21 +1,23 @@
-resource "helm_release" "network_operator" {
-  name       = "network-operator"
-  repository = var.helm_repository
-  chart      = "network-operator"
-  namespace  = "network-operator"
-  atomic     = true
-  timeout    = 600
-
-  create_namespace = true
-  version          = var.helm_version
-
-  set {
-    name  = "operator.resources.limits.cpu"
-    value = var.limit_cpu
+resource "nebius_applications_v1alpha1_k8s_release" "network-operator" {
+  lifecycle {
+    ignore_changes = all
   }
+  cluster_id = var.cluster_id
+  parent_id  = var.parent_id
 
-  set {
-    name  = "operator.resources.limits.memory"
-    value = var.limit_memory
-  }
+  application_name = "network-operator"
+  namespace        = "network-operator"
+  product_slug     = var.product_slug
+
+  #  set = {
+  #    "operator.resources.limits.cpu" : var.limit_cpu,
+  #    "operator.resources.limits.memory" : var.limit_memory
+  #  }
+  values = <<EOT
+  operator:
+    resources:
+      limits:
+        cpu: ${var.limit_cpu}
+        memory: ${var.limit_memory}
+  EOT
 }

modules/network-operator/variables.tf

@@ -1,13 +1,17 @@
-variable "helm_repository" {
-  description = "Network Operator Helm chart source repository."
+variable "cluster_id" {
+  description = "K8s cluster id."
   type        = string
-  default     = "oci://cr.eu-north1.nebius.cloud/marketplace/nebius/nvidia-network-operator/chart"
 }
 
-variable "helm_version" {
-  description = "Version of Network Operator Helm chart."
+variable "parent_id" {
+  description = "Project id."
   type        = string
-  default     = "24.4.0"
+}
+
+variable "product_slug" {
+  description = "Marketplace product slug."
+  type        = string
+  default     = "nebius/nvidia-network-operator"
 }
 
 variable "limit_cpu" {

modules/network-operator/versions.tf

@@ -0,0 +1,7 @@
+terraform {
+  required_providers {
+    nebius = {
+      source = "terraform-provider-nebius.storage.ai.nebius.cloud/nebius/nebius"
+    }
+  }
+}