Merge pull request #7571 from zalando-incubator/kube-1.30-to-alpha

kube-1.30 to alpha
zalando-incubator · May 27, 2024 · 08fbcb2 · 08fbcb2
2 parents bd33306 + 9ccd2ea
commit 08fbcb2
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 1 deletion.
diff --git a/cluster/config-defaults.yaml b/cluster/config-defaults.yaml
@@ -387,6 +387,10 @@ flannel_memory: "100Mi"
 nvidia_device_plugin_cpu: "10m"
 nvidia_device_plugin_memory: "50Mi"
 
+# nvidia dcgm exporter
+nvidia_dcgm_exporter_cpu: "10m"
+nvidia_dcgm_exporter_memory: "200Mi"
+
 # static egress controller settings
 static_egress_controller_enabled: "true"
 

diff --git a/cluster/manifests/nvidia/nvidia-gpu-device-plugin.yaml b/cluster/manifests/nvidia/nvidia-gpu-device-plugin.yaml
@@ -24,6 +24,9 @@ spec:
         component: nvidia-gpu-device-plugin
       annotations:
         logging/destination: "{{.Cluster.ConfigItems.log_destination_infra}}"
+        prometheus.io/path: /metrics
+        prometheus.io/port: "9400"
+        prometheus.io/scrape: "true"
     spec:
       serviceAccountName: nvidia
       tolerations:
@@ -48,6 +51,9 @@ spec:
       - name: device-plugin
         hostPath:
           path: /var/lib/kubelet/device-plugins
+      - name: pod-gpu-resources
+        hostPath:
+          path: /opt/podruntime/kubelet/pod-resources
       containers:
       - name: nvidia-gpu-device-plugin
         image: container-registry.zalando.net/teapot/nvidia-gpu-device-plugin:v0.14.5-master-10
@@ -68,3 +74,27 @@ spec:
         volumeMounts:
         - name: device-plugin
           mountPath: /var/lib/kubelet/device-plugins
+      - name: dcgm-exporter
+        image: container-registry.zalando.net/teapot/nvidia-dcgm-exporter:v3.3.6-3.4.2-ubuntu22.04-master-11
+        args:
+        - --kubernetes
+        - --address=:9400
+        resources:
+          requests:
+            cpu: "{{ .Cluster.ConfigItems.nvidia_dcgm_exporter_cpu }}"
+            memory: "{{ .Cluster.ConfigItems.nvidia_dcgm_exporter_memory }}"
+          limits:
+            cpu: "{{ .Cluster.ConfigItems.nvidia_dcgm_exporter_cpu }}"
+            memory: "{{ .Cluster.ConfigItems.nvidia_dcgm_exporter_memory }}"
+        ports:
+        - name: metrics
+          containerPort: 9400
+        securityContext:
+          runAsNonRoot: false
+          runAsUser: 0
+          capabilities:
+            add: ["SYS_ADMIN"]
+        volumeMounts:
+        - name: pod-gpu-resources
+          mountPath: /var/lib/kubelet/pod-resources
+          readOnly: true
diff --git a/cluster/manifests/skipper/deployment.yaml b/cluster/manifests/skipper/deployment.yaml
@@ -1,5 +1,5 @@
 {{ $internal_version := "v0.21.91-921" }}
-{{ $canary_internal_version := "v0.21.91-921" }}
+{{ $canary_internal_version := "v0.21.99-931" }}
 
 {{/* Optional canary arguments separated by "[cf724afc]" to allow whitespaces, e.g. "-foo=has a whitespace[cf724afc]-baz=qux" */}}
 {{ $canary_args := "" }}