Various updates

Author: OguzPastirmaci

manifests/node-problem-detector/values.yaml

@@ -17,4 +17,4 @@ spec:
     solvers:
     - http01:
         ingress:
-          ingressClassName: nginx
+          ingressClassName: nginx

terraform/files/cert-manager/cluster-issuer.yaml renamed to terraform/files/cert-manager/cluster-issuer-prod.yaml

@@ -0,0 +1,20 @@
+---
+apiVersion: cert-manager.io/v1
+kind: ClusterIssuer
+metadata:
+  name: le-clusterissuer
+spec:
+  acme:
+    # You must replace this email address with your own.
+    # Let's Encrypt will use this to contact you about expiring
+    # certificates, and issues related to your account.
+    email: [email protected]
+    server: https://acme-staging-v02.api.letsencrypt.org/directory
+    privateKeySecretRef:
+      # Secret resource that will be used to store the account's private key.
+      name: le-clusterissuer-secret
+    # Add a single challenge solver, HTTP01 using nginx
+    solvers:
+    - http01:
+        ingress:
+          ingressClassName: nginx

terraform/files/cert-manager/cluster-issuer-staging.yaml

@@ -520,8 +520,8 @@ logDir:
   pod: ""
 
 image:
-  repository: iad.ocir.io/hpc_limited_availability/oke-npd
-  tag: v0.8.21-1
+  repository: iad.ocir.io/idxzjcdglx2s/oke-npd
+  tag: v1.34.0-1
   # image.digest -- the image digest. If given it takes precedence over a given tag.
   digest: ""
   pullPolicy: IfNotPresent
@@ -591,6 +591,8 @@ affinity:
               - VM.GPU.A100.80G.1
               - BM.GPU.H200.8
               - BM.GPU.B200.8
+              - BM.GPU.GB200.4
+              - BM.GPU.GB200-v2.4
 
 nodeSelector: {}
 

terraform/files/node-problem-detector/values.yaml

@@ -1,9 +1,9 @@
 apiVersion: v2
 name: dcgm-exporter
 description: A Helm chart for DCGM exporter
-version: "4.5.2"
+version: "4.6.0"
 kubeVersion: ">= 1.19.0-0"
-appVersion: "4.5.2"
+appVersion: "4.6.0"
 sources:
 - https://github.com/nvidia/dcgm-exporter
 home: https://github.com/nvidia/dcgm-exporter/

terraform/files/nvidia-dcgm-exporter/Chart.yaml

@@ -46,6 +46,7 @@ spec:
       runtimeClassName: {{ .Values.runtimeClassName }}
       {{- end }}
       priorityClassName: {{ .Values.priorityClassName | default "system-node-critical" }}
+      hostPID: {{ .Values.hostPID | default false }}
       {{- if .Values.hostNetwork }}
       hostNetwork: {{ .Values.hostNetwork }}
       dnsPolicy: ClusterFirstWithHostNet
@@ -138,6 +139,10 @@ spec:
         - name: "DCGM_EXPORTER_KUBERNETES_ENABLE_POD_UID"
           value: "true"
         {{- end }}
+        {{- if .Values.kubernetes.podLabelAllowlistRegex }}
+        - name: "DCGM_EXPORTER_KUBERNETES_POD_LABEL_ALLOWLIST_REGEX"
+          value: {{ .Values.kubernetes.podLabelAllowlistRegex | join "," | quote }}
+        {{- end }}
         - name: "DCGM_EXPORTER_LISTEN"
           value: "{{ .Values.service.address }}"
         - name: NODE_NAME

terraform/files/nvidia-dcgm-exporter/templates/daemonset.yaml

@@ -17,7 +17,7 @@ image:
   pullPolicy: IfNotPresent
   # Image tag defaults to AppVersion, but you can use the tag key
   # for the image tag, e.g:
-  tag: 4.4.1-4.5.2-ubuntu22.04
+  tag: 4.4.1-4.6.0-ubuntu22.04
 
 # Change the following reference to "/etc/dcgm-exporter/default-counters.csv"
 # to stop profiling metrics from DCGM
@@ -49,6 +49,10 @@ fullnameOverride: ""
 # Overrides the deployment namespace
 namespaceOverride: ""
 
+# hostPID allows the DCGM-Exporter container to see processes on the host node
+# Default: false
+hostPID: false
+
 # Defines the runtime class that will be used by the pod
 runtimeClassName: ""
 # Defines serviceAccount names for components.
@@ -211,6 +215,16 @@ kubernetes:
   # This requires cluster-level read permissions to pods
   enablePodUID: false
 
+  # Pod label filtering configuration
+  # Filter which pod labels are included in metrics using regex patterns
+  # Empty list means all labels are included (default behavior)
+  # Examples:
+  #   podLabelAllowlistRegex:
+  #     - "^app$"                         # Exact match for "app" label
+  #     - "^app\\.kubernetes\\.io/.*"     # All labels starting with app.kubernetes.io/
+  #     - "^(tier|environment|version)$"  # Match tier, environment, or version labels
+  podLabelAllowlistRegex: []
+
   # RBAC settings for Kubernetes integration
   rbac:
     # Automatically creates ClusterRole and ClusterRoleBinding for pod access when enablePodLabels or enablePodUID is true
@@ -321,4 +335,4 @@ livenessProbe:
   periodSeconds: 5
 
 readinessProbe:
-  initialDelaySeconds: 45
+  initialDelaySeconds: 45

terraform/files/nvidia-dcgm-exporter/values.yaml

@@ -124,9 +124,13 @@ data:
         if response.status_code == 200:
             short_url = response.json().get("url") 
             parsed_short_url = urlparse(short_url)
-            if "sslip.io" in short_url:
+            netloc = parsed_url.netloc.split(":")[0]
+            ip_only = False
+            match = re.match(r"^(\d{1,3}\.?){4}$", netloc)
+            if match:
+                ip_only = True
+            if not ip_only:
                 scheme = "https"
-                netloc = parsed_url.netloc.split(":")[0]
                 return parsed_short_url.geturl().replace(f'{parsed_short_url.scheme}://{parsed_url.netloc}/', f"{scheme}://{netloc}/")
             return short_url
         else:

terraform/files/oke-ons-webhook/templates/configmap.yaml

@@ -152,7 +152,6 @@ locals {
 module "oke" {
   source  = "oracle-terraform-modules/oke/oci"
   version = "5.3.3"
-  # source    = "github.com/oracle-terraform-modules/terraform-oci-oke.git?ref=ca3e66f"
 
   providers = { oci.home = oci.home }
 

terraform/oke-cluster.tf

@@ -64,7 +64,7 @@ output "grafana_fetch_endpoint_command" {
 output "grafana_url" {
   value = var.install_node_problem_detector_kube_prometheus_stack ? (
     var.preferred_kubernetes_services == "public" ?
-    format("https://grafana.%s.sslip.io", try(data.kubernetes_service.nginx_lb[0].status[0].load_balancer[0].ingress[0].ip, try(data.oci_load_balancer_load_balancers.lbs[0].load_balancers[0].ip_addresses[0], "N/A"))) :
+    format("https://grafana.%s.%s", try(data.kubernetes_service.nginx_lb[0].status[0].load_balancer[0].ingress[0].ip, try(data.oci_load_balancer_load_balancers.lbs[0].load_balancers[0].ip_addresses[0], "N/A")), var.wildcard_dns_domain):
     format("http://%s", try(data.kubernetes_service.grafana_internal_ip[0].status[0].load_balancer[0].ingress[0].ip, try(data.oci_load_balancer_load_balancers.lbs[0].load_balancers[0].ip_addresses[0], try(data.oci_load_balancer_load_balancers.internal_lbs[0].load_balancers[0].ip_addresses[0], "N/A"))))
   ) : "N/A"
 }