Add parameter to configure health endpoint port

Signed-off-by: Tobias Giese <tgiese@nvidia.com>
kubernetes-sigs · Sep 24, 2024 · ed0ca13 · ed0ca13
1 parent d858501
commit ed0ca13
Show file tree

Hide file tree

Showing 9 changed files with 40 additions and 22 deletions.
diff --git a/cmd/nfd-master/main.go b/cmd/nfd-master/main.go
@@ -33,8 +33,7 @@ import (
 
 const (
 	// ProgramName is the canonical name of this program
-	ProgramName    = "nfd-master"
-	GrpcHealthPort = 8082
+	ProgramName = "nfd-master"
 )
 
 func main() {
@@ -108,7 +107,6 @@ func main() {
 	utils.ConfigureGrpcKlog()
 
 	// Get new NfdMaster instance
-	args.GrpcHealthPort = GrpcHealthPort
 	instance, err := master.NewNfdMaster(master.WithArgs(args))
 	if err != nil {
 		klog.ErrorS(err, "failed to initialize NfdMaster instance")
@@ -149,6 +147,8 @@ func initFlags(flagset *flag.FlagSet) (*master.Args, *master.ConfigOverrideArgs)
 			" DEPRECATED: will be removed in a future release along with the deprecated gRPC API.")
 	flagset.IntVar(&args.MetricsPort, "metrics", 8081,
 		"Port on which to expose metrics.")
+	flagset.IntVar(&args.GrpcHealthPort, "grpc-health", 8082,
+		"Port on which to expose the grpc health endpoint.")
 	flagset.BoolVar(&args.Prune, "prune", false,
 		"Prune all NFD related attributes from all nodes of the cluster and exit.")
 	flagset.BoolVar(&args.VerifyNodeName, "verify-node-name", false,

diff --git a/cmd/nfd-topology-updater/main.go b/cmd/nfd-topology-updater/main.go
@@ -38,7 +38,6 @@ const (
 	// ProgramName is the canonical name of this program
 	ProgramName       = "nfd-topology-updater"
 	kubeletSecurePort = 10250
-	GrpcHealthPort    = 8082
 )
 
 var DefaultKubeletStateDir = path.Join(string(hostpath.VarDir), "lib", "kubelet")
@@ -57,7 +56,6 @@ func main() {
 	utils.ConfigureGrpcKlog()
 
 	// Get new TopologyUpdater instance
-	args.GrpcHealthPort = GrpcHealthPort
 	instance, err := topology.NewTopologyUpdater(*args, *resourcemonitorArgs)
 	if err != nil {
 		klog.ErrorS(err, "failed to initialize topology updater instance")
@@ -115,6 +113,8 @@ func initFlags(flagset *flag.FlagSet) (*topology.Args, *resourcemonitor.Args) {
 		"Kube config file.")
 	flagset.IntVar(&args.MetricsPort, "metrics", 8081,
 		"Port on which to expose metrics.")
+	flagset.IntVar(&args.GrpcHealthPort, "grpc-health", 8082,
+		"Port on which to expose the grpc health endpoint.")
 	flagset.DurationVar(&resourcemonitorArgs.SleepInterval, "sleep-interval", time.Duration(60)*time.Second,
 		"Time to sleep between CR updates. zero means no CR updates on interval basis. [Default: 60s]")
 	flagset.StringVar(&resourcemonitorArgs.Namespace, "watch-namespace", "*",

diff --git a/cmd/nfd-worker/main.go b/cmd/nfd-worker/main.go
@@ -32,8 +32,7 @@ import (
 
 const (
 	// ProgramName is the canonical name of this program
-	ProgramName    = "nfd-worker"
-	GrpcHealthPort = 8082
+	ProgramName = "nfd-worker"
 )
 
 func main() {
@@ -80,7 +79,6 @@ func main() {
 	utils.ConfigureGrpcKlog()
 
 	// Get new NfdWorker instance
-	args.GrpcHealthPort = GrpcHealthPort
 	instance, err := worker.NewNfdWorker(worker.WithArgs(args))
 	if err != nil {
 		klog.ErrorS(err, "failed to initialize NfdWorker instance")
@@ -138,6 +136,8 @@ func initFlags(flagset *flag.FlagSet) (*worker.Args, *worker.ConfigOverrideArgs)
 		"Do not publish feature labels")
 	flagset.IntVar(&args.MetricsPort, "metrics", 8081,
 		"Port on which to expose metrics.")
+	flagset.IntVar(&args.GrpcHealthPort, "grpc-health", 8082,
+		"Port on which to expose the grpc health endpoint.")
 	flagset.StringVar(&args.Options, "options", "",
 		"Specify config options from command line. Config options are specified "+
 			"in the same format as in the config file (i.e. json or yaml). These options")

diff --git a/deployment/helm/node-feature-discovery/templates/master.yaml b/deployment/helm/node-feature-discovery/templates/master.yaml
@@ -116,6 +116,7 @@ spec:
             - "-feature-gates={{ $key }}={{ $value }}"
             {{- end }}
             - "-metrics={{ .Values.master.metricsPort  | default "8081" }}"
+            - "-grpc-health={{ .Values.master.healthPort  | default "8082" }}"
             {{- with .Values.master.extraArgs }}
             {{- toYaml . | nindent 12 }}
             {{- end }}

diff --git a/deployment/helm/node-feature-discovery/templates/topologyupdater.yaml b/deployment/helm/node-feature-discovery/templates/topologyupdater.yaml
@@ -89,7 +89,8 @@ spec:
           # Disable kubelet state tracking by giving an empty path
           - "-kubelet-state-dir="
           {{- end }}
-          - -metrics={{ .Values.topologyUpdater.metricsPort | default "8081"}}
+          - "-metrics={{ .Values.topologyUpdater.metricsPort | default "8081"}}"
+          - "-grpc-health={{ .Values.topologyUpdater.healthPort  | default "8082" }}"
           {{- with .Values.topologyUpdater.extraArgs }}
           {{- toYaml . | nindent 10 }}
           {{- end }}

diff --git a/deployment/helm/node-feature-discovery/templates/worker.yaml b/deployment/helm/node-feature-discovery/templates/worker.yaml
@@ -71,19 +71,20 @@ spec:
         command:
         - "nfd-worker"
         args:
-{{- if not .Values.featureGates.NodeFeatureAPI }}
+        {{- if not .Values.featureGates.NodeFeatureAPI }}
         - "-server={{ include "node-feature-discovery.fullname" . }}-master:{{ .Values.master.service.port }}"
-{{- end }}
-{{- if .Values.tls.enable }}
+        {{- end }}
+        {{- if .Values.tls.enable }}
         - "-ca-file=/etc/kubernetes/node-feature-discovery/certs/ca.crt"
         - "-key-file=/etc/kubernetes/node-feature-discovery/certs/tls.key"
         - "-cert-file=/etc/kubernetes/node-feature-discovery/certs/tls.crt"
-{{- end }}
-# Go over featureGate and add the feature-gate flag
-{{- range $key, $value := .Values.featureGates }}
+        {{- end }}
+        # Go over featureGate and add the feature-gate flag
+        {{- range $key, $value := .Values.featureGates }}
         - "-feature-gates={{ $key }}={{ $value }}"
-{{- end }}
+        {{- end }}
         - "-metrics={{ .Values.worker.metricsPort | default "8081"}}"
+        - "-grpc-health={{ .Values.worker.healthPort  | default "8082" }}"
         {{- with .Values.gc.extraArgs }}
         {{- toYaml . | nindent 8 }}
         {{- end }}
@@ -125,11 +126,11 @@ spec:
         - name: nfd-worker-conf
           mountPath: "/etc/kubernetes/node-feature-discovery"
           readOnly: true
-{{- if .Values.tls.enable }}
+        {{- if .Values.tls.enable }}
         - name: nfd-worker-cert
           mountPath: "/etc/kubernetes/node-feature-discovery/certs"
           readOnly: true
-{{- end }}
+        {{- end }}
       volumes:
         - name: host-boot
           hostPath:
@@ -166,12 +167,12 @@ spec:
             items:
               - key: nfd-worker.conf
                 path: nfd-worker.conf
-{{- if .Values.tls.enable }}
+        {{- if .Values.tls.enable }}
         - name: nfd-worker-cert
           secret:
             secretName: nfd-worker-cert
-{{- end }}
-    {{- with .Values.worker.nodeSelector }}
+        {{- end }}
+      {{- with .Values.worker.nodeSelector }}
       nodeSelector:
         {{- toYaml . | nindent 8 }}
       {{- end }}

diff --git a/deployment/helm/node-feature-discovery/values.yaml b/deployment/helm/node-feature-discovery/values.yaml
@@ -58,6 +58,7 @@ master:
   # be removed with it in a future release
   port: 8080
   metricsPort: 8081
+  healthPort: 8082
   instance:
   featureApi:
   resyncPeriod:
@@ -406,6 +407,7 @@ worker:
 ### <NFD-WORKER-CONF-END-DO-NOT-REMOVE>
 
   metricsPort: 8081
+  healthPort: 8082
   daemonsetAnnotations: {}
   podSecurityContext: {}
     # fsGroup: 2000
@@ -497,6 +499,7 @@ topologyUpdater:
     create: true
 
   metricsPort: 8081
+  healthPort: 8082
   kubeletConfigPath:
   kubeletPodResourcesSockPath:
   updateInterval: 60s

diff --git a/docs/deployment/helm.md b/docs/deployment/helm.md
@@ -180,6 +180,7 @@ API's you need to install the prometheus operator in your cluster.
 | `master.hostNetwork`                | bool    | false                            | Specifies whether to enable or disable running the container in the host's network namespace                                                                                                           |
 | `master.port`                       | integer |                                  | Specifies the TCP port that nfd-master listens for incoming requests. **NOTE**: this parameter is related to the deprecated gRPC API and will be removed with it in a future release                   |
 | `master.metricsPort`                | integer | 8081                             | Port on which to expose metrics from components to prometheus operator                                                                                                                                 |
+| `master.healthPort`                 | integer | 8082                             | Port on which to expose the grpc health endpoint                                                                                                                                                       |
 | `master.instance`                   | string  |                                  | Instance name. Used to separate annotation namespaces for multiple parallel deployments                                                                                                                |
 | `master.resyncPeriod`               | string  |                                  | NFD API controller resync period.                                                                                                                                                                      |
 | `master.extraLabelNs`               | array   | []                               | List of allowed extra label namespaces                                                                                                                                                                 |
@@ -217,7 +218,8 @@ API's you need to install the prometheus operator in your cluster.
 | `worker.*`                          | dict   |                         | NFD worker daemonset configuration                                                                                                                                                                   |
 | `worker.enable`                     | bool   | true                    | Specifies whether nfd-worker should be deployed                                                                                                                                                      |
 | `worker.hostNetwork`                | bool   | false                   | Specifies whether to enable or disable running the container in the host's network namespace                                                                                                         |
-| `worker.metricsPort*`               | int    | 8081                    | Port on which to expose metrics from components to prometheus operator                                                                                                                               |
+| `worker.metricsPort`                | int    | 8081                    | Port on which to expose metrics from components to prometheus operator                                                                                                                               |
+| `worker.healthPort`                 | int    | 8082                    | Port on which to expose the grpc health endpoint                                                                                                                                                     |
 | `worker.config`                     | dict   |                         | NFD worker [configuration](../reference/worker-configuration-reference)                                                                                                                              |
 | `worker.podSecurityContext`         | dict   | {}                      | [PodSecurityContext](https://kubernetes.io/docs/tasks/configure-pod-container/security-context/#set-the-security-context-for-a-pod) holds pod-level security attributes and common container settins |
 | `worker.securityContext`            | dict   | {}                      | Container [security settings](https://kubernetes.io/docs/tasks/configure-pod-container/security-context/#set-the-security-context-for-a-container)                                                   |
@@ -252,6 +254,7 @@ API's you need to install the prometheus operator in your cluster.
 | `topologyUpdater.serviceAccount.name`         | string  |                          | The name of the service account for topology updater to use. If not set and create is true, a name is generated using the fullname template and `-topology-updater` suffix                       |
 | `topologyUpdater.rbac.create`                 | bool    | true                     | Specifies whether to create [RBAC][rbac] configuration for topology updater                                                                                                                      |
 | `topologyUpdater.metricsPort`                 | integer | 8081                     | Port on which to expose prometheus metrics                                                                                                                                                       |
+| `topologyUpdater.healthPort`                  | integer | 8082                     | Port on which to expose the grpc health endpoint                                                                                                                                                 |
 | `topologyUpdater.kubeletConfigPath`           | string  | ""                       | Specifies the kubelet config host path                                                                                                                                                           |
 | `topologyUpdater.kubeletPodResourcesSockPath` | string  | ""                       | Specifies the kubelet sock path to read pod resources                                                                                                                                            |
 | `topologyUpdater.updateInterval`              | string  | 60s                      | Time to sleep between CR updates. Non-positive value implies no CR update.                                                                                                                       |

diff --git a/node-feature-discovery.iml b/node-feature-discovery.iml
@@ -0,0 +1,9 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="WEB_MODULE" version="4">
+  <component name="NewModuleRootManager" inherit-compiler-output="true">
+    <exclude-output />
+    <content url="file://$MODULE_DIR$" />
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+</module>