Add logVerbosity Helm chart parameter

Signed-off-by: Dr. Jan-Philip Gehrcke <[email protected]>

Author: jgehrcke

deployments/helm/nvidia-dra-driver-gpu/templates/controller.yaml

@@ -52,7 +52,7 @@ spec:
           {{- toYaml .Values.controller.containers.computeDomain.securityContext | nindent 10 }}
         image: {{ include "nvidia-dra-driver-gpu.fullimage" . }}
         imagePullPolicy: {{ .Values.image.pullPolicy }}
-        command: ["compute-domain-controller", "-v", "6"]
+        command: ["compute-domain-controller", "-v", "{{ .Values.logVerbosity }}"]
         resources:
           {{- toYaml .Values.controller.containers.computeDomain.resources | nindent 10 }}
         env:

deployments/helm/nvidia-dra-driver-gpu/templates/kubeletplugin.yaml

@@ -99,7 +99,7 @@ spec:
             sed -i 's/^ModifyDeviceFiles: 1$/ModifyDeviceFiles: 0/' root/gpu-params
             mount --bind root/gpu-params /proc/driver/nvidia/params
           fi
-          compute-domain-kubelet-plugin -v 6
+          compute-domain-kubelet-plugin -v {{ .Values.logVerbosity }}
         resources:
           {{- toYaml .Values.kubeletPlugin.containers.computeDomains.resources | nindent 10 }}
         {{/*
@@ -181,7 +181,7 @@ spec:
             sed -i 's/^ModifyDeviceFiles: 1$/ModifyDeviceFiles: 0/' root/gpu-params
             mount --bind root/gpu-params /proc/driver/nvidia/params
           fi
-          gpu-kubelet-plugin -v 6
+          gpu-kubelet-plugin -v {{ .Values.logVerbosity }}
         resources:
           {{- toYaml .Values.kubeletPlugin.containers.gpus.resources | nindent 10 }}
         {{/*

deployments/helm/nvidia-dra-driver-gpu/values.yaml

@@ -69,6 +69,30 @@ resources:
 #   LoggingBetaOptions: true           # Kubernetes logging beta features
 featureGates: {}
 
+# klog verbosity for all components. Zero or greater. Higher number: higher
+# verbosity. Warning and error log messages are always logged (regardless of
+# this setting). Info-level messages on level 0 and are also always logged (this
+# cannot be prevented).
+#
+# Level 0:
+# - Configuration (startup)
+# - Permanent error upon (un)preparing device
+#
+# Level 1 (default):
+# - Device (un)prepare confirmation with claim UID
+# - Workqueue reconciliation failures (noisy: mainly expected, retryable errors)
+# - Controller: cleanup of stale objects
+#
+# Level 2:
+# - Unprepare noop
+# - Controller: added/updated object callback confirmation
+#
+# Level 6 and higher:
+# - Checkpoint file updates
+# - Kubelet plugin GRPC request/response detail
+#
+logVerbosity: "1"
+
 # Webhook configuration
 webhook:
   enabled: false