sysdump: collect 'clustermesh-apiserver' + improvements

internal/cli/cmd/sysdump.go

@@ -52,6 +52,12 @@ func newCmdSysdump() *cobra.Command {
 	cmd.Flags().StringVar(&sysdumpOptions.CiliumOperatorNamespace,
 		"cilium-operator-namespace", sysdump.DefaultCiliumOperatorNamespace,
 		"The namespace Cilium operator is running in")
+	cmd.Flags().StringVar(&sysdumpOptions.ClustermeshApiserverLabelSelector,
+		"clustermesh-apiserver-label-selector", sysdump.DefaultClustermeshApiserverLabelSelector,
+		"The labels used to target 'clustermesh-apiserver' pods")
+	cmd.Flags().StringVar(&sysdumpOptions.ClustermeshApiserverNamespace,
+		"clustermesh-apiserver-namespace", sysdump.DefaultClustermeshApiserverNamespace,
+		"The namespace Cilium operator is running in")
 	cmd.Flags().BoolVar(&sysdumpOptions.Debug,
 		"debug", sysdump.DefaultDebug,
 		"Whether to enable debug logging")

sysdump/constants.go

@@ -10,19 +10,20 @@ import (
 )
 
 const (
-	awsNodeDaemonSetName         = "aws-node"
-	awsNodeDaemonSetNamespace    = "kube-system"
-	ciliumAgentContainerName     = "cilium-agent"
-	ciliumConfigConfigMapName    = "cilium-config"
-	ciliumDaemonSetName          = "cilium"
-	ciliumEtcdSecretsSecretName  = "cilium-etcd-secrets"
-	ciliumOperatorDeploymentName = "cilium-operator"
-	hubbleContainerName          = "hubble"
-	hubbleDaemonSetName          = "hubble"
-	hubbleRelayContainerName     = "hubble-relay"
-	hubbleRelayDeploymentName    = "hubble-relay"
-	hubbleUIDeploymentName       = "hubble-ui"
-	redacted                     = "XXXXXX"
+	awsNodeDaemonSetName               = "aws-node"
+	awsNodeDaemonSetNamespace          = "kube-system"
+	ciliumAgentContainerName           = "cilium-agent"
+	ciliumConfigConfigMapName          = "cilium-config"
+	ciliumDaemonSetName                = "cilium"
+	ciliumEtcdSecretsSecretName        = "cilium-etcd-secrets"
+	ciliumOperatorDeploymentName       = "cilium-operator"
+	clustermeshApiserverDeploymentName = "clustermesh-apiserver"
+	hubbleContainerName                = "hubble"
+	hubbleDaemonSetName                = "hubble"
+	hubbleRelayContainerName           = "hubble-relay"
+	hubbleRelayDeploymentName          = "hubble-relay"
+	hubbleUIDeploymentName             = "hubble-ui"
+	redacted                           = "XXXXXX"
 )
 
 const (
@@ -39,6 +40,7 @@ const (
 	ciliumNetworkPoliciesFileName            = "ciliumnetworkpolicies-<ts>.yaml"
 	ciliumNodesFileName                      = "ciliumnodes-<ts>.yaml"
 	ciliumOperatorDeploymentFileName         = "cilium-operator-deployment-<ts>.yaml"
+	clustermeshApiserverDeploymentFileName   = "clustermesh-apiserver-deployment-<ts>.yaml"
 	eniconfigsFileName                       = "aws-eniconfigs-<ts>.yaml"
 	gopsFileName                             = "gops-%s-%s-<ts>-%s.txt"
 	hubbleDaemonsetFileName                  = "hubble-daemonset-<ts>.yaml"

sysdump/defaults.go

@@ -14,26 +14,28 @@ const (
 )
 
 const (
-	DefaultCiliumLabelSelector         = labelPrefix + "cilium"
-	DefaultCiliumNamespace             = "kube-system"
-	DefaultCiliumOperatorLabelSelector = "io.cilium/app=operator"
-	DefaultCiliumOperatorNamespace     = DefaultCiliumNamespace
-	DefaultDebug                       = false
-	DefaultHubbleLabelSelector         = labelPrefix + "hubble"
-	DefaultHubbleNamespace             = DefaultCiliumNamespace
-	DefaultHubbleFlowsCount            = 10000
-	DefaultHubbleFlowsTimeout          = 5 * time.Second
-	DefaultHubbleRelayLabelSelector    = labelPrefix + "hubble-relay"
-	DefaultHubbleRelayNamespace        = DefaultCiliumNamespace
-	DefaultHubbleUILabelSelector       = labelPrefix + "hubble-ui"
-	DefaultHubbleUINamespace           = DefaultCiliumNamespace
-	DefaultLargeSysdumpAbortTimeout    = 5 * time.Second
-	DefaultLargeSysdumpThreshold       = 20
-	DefaultLogsSinceTime               = 8760 * time.Hour // 1y
-	DefaultLogsLimitBytes              = 1073741824       // 1GiB
-	DefaultNodeList                    = ""
-	DefaultQuick                       = false
-	DefaultOutputFileName              = "cilium-sysdump-<ts>" // "<ts>" will be replaced with the timestamp
+	DefaultCiliumLabelSelector               = labelPrefix + "cilium"
+	DefaultCiliumNamespace                   = "kube-system"
+	DefaultCiliumOperatorLabelSelector       = "io.cilium/app=operator"
+	DefaultCiliumOperatorNamespace           = DefaultCiliumNamespace
+	DefaultClustermeshApiserverLabelSelector = "k8s-app=clustermesh-apiserver"
+	DefaultClustermeshApiserverLabelSelector = labelPrefix + "clustermesh-apiserver"
+	DefaultDebug                             = false
+	DefaultHubbleLabelSelector               = labelPrefix + "hubble"
+	DefaultHubbleNamespace                   = DefaultCiliumNamespace
+	DefaultHubbleFlowsCount                  = 10000
+	DefaultHubbleFlowsTimeout                = 5 * time.Second
+	DefaultHubbleRelayLabelSelector          = labelPrefix + "hubble-relay"
+	DefaultHubbleRelayNamespace              = DefaultCiliumNamespace
+	DefaultHubbleUILabelSelector             = labelPrefix + "hubble-ui"
+	DefaultHubbleUINamespace                 = DefaultCiliumNamespace
+	DefaultLargeSysdumpAbortTimeout          = 5 * time.Second
+	DefaultLargeSysdumpThreshold             = 20
+	DefaultLogsSinceTime                     = 8760 * time.Hour // 1y
+	DefaultLogsLimitBytes                    = 1073741824       // 1GiB
+	DefaultNodeList                          = ""
+	DefaultQuick                             = false
+	DefaultOutputFileName                    = "cilium-sysdump-<ts>" // "<ts>" will be replaced with the timestamp
 )
 
 var (

sysdump/sysdump.go

@@ -34,6 +34,10 @@ type Options struct {
 	CiliumOperatorLabelSelector string
 	// The namespace Cilium operator is running in.
 	CiliumOperatorNamespace string
+	// The labels used to target 'clustermesh-apiserver' pods.
+	ClustermeshApiserverLabelSelector string
+	// The namespace 'clustermesh-apiserver' is running in.
+	ClustermeshApiserverNamespace string
 	// Whether to enable debug logging.
 	Debug bool
 	// The labels used to target Hubble pods.
@@ -448,7 +452,7 @@ func (c *Collector) Run() error {
 			Description: "Collecting the Cilium operator deployment",
 			Quick:       true,
 			Task: func(ctx context.Context) error {
-				v, err := c.client.GetDeployment(ctx, c.options.CiliumNamespace, ciliumOperatorDeploymentName, metav1.GetOptions{})
+				v, err := c.client.GetDeployment(ctx, c.options.CiliumOperatorNamespace, ciliumOperatorDeploymentName, metav1.GetOptions{})
 				if err != nil {
 					return fmt.Errorf("failed to collect the Cilium operator deployment: %w", err)
 				}
@@ -458,6 +462,24 @@ func (c *Collector) Run() error {
 				return nil
 			},
 		},
+		{
+			Description: "Collecting the 'clustermesh-apiserver' deployment",
+			Quick:       true,
+			Task: func(ctx context.Context) error {
+				v, err := c.client.GetDeployment(ctx, c.options.ClustermeshApiserverNamespace, clustermeshApiserverDeploymentName, metav1.GetOptions{})
+				if err != nil {
+					if errors.IsNotFound(err) {
+						c.logWarn("deployment %q not found in namespace %q - this is expected if 'clustermesh-apiserver' isn't enabled", clustermeshApiserverDeploymentName, c.options.ClustermeshApiserverNamespace)
+						return nil
+					}
+					return fmt.Errorf("failed to collect the 'clustermesh-apiserver' deployment: %w", err)
+				}
+				if err := writeYaml(absoluteTempPath(clustermeshApiserverDeploymentFileName), v); err != nil {
+					return fmt.Errorf("failed to collect the 'clustermesh-apiserver' deployment: %w", err)
+				}
+				return nil
+			},
+		},
 		{
 			CreatesSubtasks: true,
 			Description:     "Collecting gops stats from Cilium pods",
@@ -560,6 +582,23 @@ func (c *Collector) Run() error {
 				return nil
 			},
 		},
+		{
+			CreatesSubtasks: true,
+			Description:     "Collecting logs from 'clustermesh-apiserver' pods",
+			Quick:           false,
+			Task: func(ctx context.Context) error {
+				p, err := c.client.ListPods(ctx, c.options.CiliumNamespace, metav1.ListOptions{
+					LabelSelector: c.options.ClustermeshApiserverLabelSelector,
+				})
+				if err != nil {
+					return fmt.Errorf("failed to get logs from 'clustermesh-apiserver' pods")
+				}
+				if err := c.submitLogsTasks(ctx, filterPods(p, nodeList), c.options.LogsSinceTime, c.options.LogsLimitBytes, absoluteTempPath); err != nil {
+					return fmt.Errorf("failed to collect logs from 'clustermesh-apiserver' pods")
+				}
+				return nil
+			},
+		},
 		{
 			CreatesSubtasks: true,
 			Description:     "Collecting logs from Hubble pods",