chore: decouple scraping logic from vmagent code

api/operator/v1beta1/common_scrapeparams.go

@@ -8,6 +8,8 @@ import (
 
 	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/utils/ptr"
+	"sigs.k8s.io/controller-runtime/pkg/client"
 )
 
 // AttachMetadata configures metadata attachment
@@ -539,6 +541,11 @@ type CommonScrapeParams struct {
 	// it doesn't affect metrics ingested directly by push API's
 	// +optional
 	ExternalLabels map[string]string `json:"externalLabels,omitempty"`
+	// IngestOnlyMode switches vmagent into unmanaged mode
+	// it disables any config generation for scraping
+	// Currently it prevents vmagent from managing tls and auth options for remote write
+	// +optional
+	IngestOnlyMode *bool `json:"ingestOnlyMode,omitempty"`
 	// EnableKubernetesAPISelectors instructs vmagent to use CRD scrape objects spec.selectors for
 	// Kubernetes API list and watch requests.
 	// https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/#list-and-watch-filtering
@@ -548,8 +555,58 @@ type CommonScrapeParams struct {
 	CommonScrapeSecurityEnforcements `json:",inline,omitempty"`
 }
 
+func (cr *CommonScrapeParams) externalLabelName() string {
+	// Use "prometheus" external label name by default if field is missing.
+	// in case of migration from prometheus to vmagent, it helps to have same labels
+	// Do not add external label if field is set to empty string.
+	if cr.ExternalLabelName != nil {
+		return *cr.ExternalLabelName
+	}
+	if cr.VMAgentExternalLabelName != nil {
+		return *cr.VMAgentExternalLabelName
+	}
+	return "prometheus"
+}
+
+func (cr *CommonScrapeParams) externalLabels(defaultLabelValue string) map[string]string {
+	m := map[string]string{}
+
+	prometheusExternalLabelName := cr.externalLabelName()
+	if len(prometheusExternalLabelName) > 0 {
+		m[prometheusExternalLabelName] = defaultLabelValue
+	}
+
+	for n, v := range cr.ExternalLabels {
+		m[n] = v
+	}
+	return m
+}
+
+// ScrapeSelectors gets object and namespace sepectors
+func (cr *CommonScrapeParams) ScrapeSelectors(scrape client.Object) (*metav1.LabelSelector, *metav1.LabelSelector) {
+	switch s := scrape.(type) {
+	case *VMNodeScrape:
+		return cr.NodeScrapeSelector, cr.NodeScrapeNamespaceSelector
+	case *VMServiceScrape:
+		return cr.ServiceScrapeSelector, cr.ServiceScrapeNamespaceSelector
+	case *VMPodScrape:
+		return cr.PodScrapeSelector, cr.PodScrapeNamespaceSelector
+	case *VMProbe:
+		return cr.ProbeSelector, cr.ProbeNamespaceSelector
+	case *VMStaticScrape:
+		return cr.StaticScrapeSelector, cr.StaticScrapeNamespaceSelector
+	case *VMScrapeConfig:
+		return cr.ScrapeConfigSelector, cr.ScrapeConfigNamespaceSelector
+	default:
+		panic(fmt.Sprintf("BUG: scrape kind %T is not supported", s))
+	}
+}
+
 // isUnmanaged checks if object should managed any config objects
 func (cr *CommonScrapeParams) isUnmanaged() bool {
+	if ptr.Deref(cr.IngestOnlyMode, false) {
+		return true
+	}
 	return !cr.SelectAllByDefault &&
 		cr.NodeScrapeSelector == nil && cr.NodeScrapeNamespaceSelector == nil &&
 		cr.ServiceScrapeSelector == nil && cr.ServiceScrapeNamespaceSelector == nil &&
@@ -561,36 +618,54 @@ func (cr *CommonScrapeParams) isUnmanaged() bool {
 
 // isNodeScrapeUnmanaged checks if scraping agent should managed any VMNodeScrape objects
 func (cr *CommonScrapeParams) isNodeScrapeUnmanaged() bool {
+	if ptr.Deref(cr.IngestOnlyMode, false) {
+		return true
+	}
 	return !cr.SelectAllByDefault &&
 		cr.NodeScrapeSelector == nil && cr.NodeScrapeNamespaceSelector == nil
 }
 
 // isServiceScrapeUnmanaged checks if scraping agent should managed any VMServiceScrape objects
 func (cr *CommonScrapeParams) isServiceScrapeUnmanaged() bool {
+	if ptr.Deref(cr.IngestOnlyMode, false) {
+		return true
+	}
 	return !cr.SelectAllByDefault &&
 		cr.ServiceScrapeSelector == nil && cr.ServiceScrapeNamespaceSelector == nil
 }
 
 // isUnmanaged checks if scraping agent should managed any VMPodScrape objects
 func (cr *CommonScrapeParams) isPodScrapeUnmanaged() bool {
+	if ptr.Deref(cr.IngestOnlyMode, false) {
+		return true
+	}
 	return !cr.SelectAllByDefault &&
 		cr.PodScrapeSelector == nil && cr.PodScrapeNamespaceSelector == nil
 }
 
 // isProbeUnmanaged checks if scraping agent should managed any VMProbe objects
 func (cr *CommonScrapeParams) isProbeUnmanaged() bool {
+	if ptr.Deref(cr.IngestOnlyMode, false) {
+		return true
+	}
 	return !cr.SelectAllByDefault &&
 		cr.ProbeSelector == nil && cr.ProbeNamespaceSelector == nil
 }
 
 // isStaticScrapeUnmanaged checks if scraping agent should managed any VMStaticScrape objects
 func (cr *CommonScrapeParams) isStaticScrapeUnmanaged() bool {
+	if ptr.Deref(cr.IngestOnlyMode, false) {
+		return true
+	}
 	return !cr.SelectAllByDefault &&
 		cr.StaticScrapeSelector == nil && cr.StaticScrapeNamespaceSelector == nil
 }
 
 // isScrapeConfigUnmanaged checks if scraping agent should managed any VMScrapeConfig objects
 func (cr *CommonScrapeParams) isScrapeConfigUnmanaged() bool {
+	if ptr.Deref(cr.IngestOnlyMode, false) {
+		return true
+	}
 	return !cr.SelectAllByDefault &&
 		cr.ScrapeConfigSelector == nil && cr.ScrapeConfigNamespaceSelector == nil
 }

api/operator/v1beta1/vmagent_types.go

@@ -34,14 +34,12 @@ type VMAgentSpec struct {
 	// +optional
 	// +kubebuilder:validation:Enum=default;json
 	LogFormat string `json:"logFormat,omitempty"`
-
 	// APIServerConfig allows specifying a host and auth methods to access apiserver.
 	// If left empty, VMAgent is assumed to run inside of the cluster
 	// and will discover API servers automatically and use the pod's CA certificate
 	// and bearer token file at /var/run/secrets/kubernetes.io/serviceaccount/.
 	// +optional
 	APIServerConfig *APIServerConfig `json:"apiServerConfig,omitempty"`
-
 	// RemoteWrite list of victoria metrics /some other remote write system
 	// for vm it must looks like: http://victoria-metrics-single:8428/api/v1/write
 	// or for cluster different url
@@ -105,11 +103,6 @@ type VMAgentSpec struct {
 
 	// ClaimTemplates allows adding additional VolumeClaimTemplates for VMAgent in StatefulMode
 	ClaimTemplates []corev1.PersistentVolumeClaim `json:"claimTemplates,omitempty"`
-	// IngestOnlyMode switches vmagent into unmanaged mode
-	// it disables any config generation for scraping
-	// Currently it prevents vmagent from managing tls and auth options for remote write
-	// +optional
-	IngestOnlyMode bool `json:"ingestOnlyMode,omitempty"`
 
 	// License allows to configure license key to be used for enterprise features.
 	// Using license key is supported starting from VictoriaMetrics v1.94.0.
@@ -218,6 +211,11 @@ func (cr *VMAgent) GetShardCount() int {
 	return *cr.Spec.ShardCount
 }
 
+// ExternalLabels returns external labels for scraping
+func (cr *VMAgent) ExternalLabels() map[string]string {
+	return cr.Spec.externalLabels(fmt.Sprintf("%s/%s", cr.Namespace, cr.Name))
+}
+
 // GetReloadURL implements reloadable interface
 func (cr *VMAgent) GetReloadURL(host string) string {
 	return BuildLocalURL(reloadAuthKey, host, cr.Spec.Port, reloadPath, cr.Spec.ExtraArgs)
@@ -567,7 +565,7 @@ func (cr *VMAgent) IsOwnsServiceAccount() bool {
 }
 
 func (cr *VMAgent) GetClusterRoleName() string {
-	return fmt.Sprintf("monitoring:%s:vmagent-%s", cr.Namespace, cr.Name)
+	return fmt.Sprintf("monitoring:%s:%s", cr.Namespace, cr.PrefixedName())
 }
 
 // AsURL - returns url for http access
@@ -612,67 +610,36 @@ func (*VMAgent) ProbeNeedLiveness() bool {
 	return true
 }
 
-// IsUnmanaged checks if object should managed any config objects
-func (cr *VMAgent) IsUnmanaged() bool {
-	// fast path
-	if cr.Spec.IngestOnlyMode {
-		return true
-	}
-	return cr.Spec.isUnmanaged()
-}
-
-// IsNodeScrapeUnmanaged checks if vmagent should managed any VMNodeScrape objects
-func (cr *VMAgent) IsNodeScrapeUnmanaged() bool {
-	// fast path
-	if cr.Spec.IngestOnlyMode {
-		return true
-	}
-	return cr.Spec.isNodeScrapeUnmanaged()
+// ScrapeSelectors gets object and namespace sepectors
+func (cr *VMAgent) ScrapeSelectors(scrape client.Object) (*metav1.LabelSelector, *metav1.LabelSelector) {
+	return cr.Spec.ScrapeSelectors(scrape)
 }
 
-// IsServiceScrapeUnmanaged checks if vmagent should managed any VMServiceScrape objects
-func (cr *VMAgent) IsServiceScrapeUnmanaged() bool {
-	// fast path
-	if cr.Spec.IngestOnlyMode {
-		return true
-	}
-	return cr.Spec.isServiceScrapeUnmanaged()
-}
-
-// IsUnmanaged checks if vmagent should managed any VMPodScrape objects
-func (cr *VMAgent) IsPodScrapeUnmanaged() bool {
-	// fast path
-	if cr.Spec.IngestOnlyMode {
+// IsUnmanaged checks if object should managed any config objects
+func (cr *VMAgent) IsUnmanaged(scrape client.Object) bool {
+	if !cr.DeletionTimestamp.IsZero() || cr.Spec.ParsingError != "" {
 		return true
 	}
-	return cr.Spec.isPodScrapeUnmanaged()
-}
-
-// IsProbeUnmanaged checks if vmagent should managed any VMProbe objects
-func (cr *VMAgent) IsProbeUnmanaged() bool {
-	// fast path
-	if cr.Spec.IngestOnlyMode {
-		return true
+	if scrape == nil {
+		return cr.Spec.isUnmanaged()
+	}
+	switch s := scrape.(type) {
+	case *VMNodeScrape:
+		return cr.Spec.DaemonSetMode || cr.Spec.isNodeScrapeUnmanaged()
+	case *VMServiceScrape:
+		return cr.Spec.DaemonSetMode || cr.Spec.isServiceScrapeUnmanaged()
+	case *VMPodScrape:
+		return cr.Spec.isPodScrapeUnmanaged()
+	case *VMProbe:
+		return cr.Spec.DaemonSetMode || cr.Spec.isProbeUnmanaged()
+	case *VMStaticScrape:
+		return cr.Spec.DaemonSetMode || cr.Spec.isStaticScrapeUnmanaged()
+	case *VMScrapeConfig:
+		return cr.Spec.DaemonSetMode || cr.Spec.isScrapeConfigUnmanaged()
+	default:
+		panic(fmt.Sprintf("BUG: scrape kind %T is not supported", s))
 	}
-	return cr.Spec.isProbeUnmanaged()
-}
 
-// IsStaticScrapeUnmanaged checks if vmagent should managed any VMStaticScrape objects
-func (cr *VMAgent) IsStaticScrapeUnmanaged() bool {
-	// fast path
-	if cr.Spec.IngestOnlyMode {
-		return true
-	}
-	return cr.Spec.isStaticScrapeUnmanaged()
-}
-
-// IsScrapeConfigUnmanaged checks if vmagent should managed any VMScrapeConfig objects
-func (cr *VMAgent) IsScrapeConfigUnmanaged() bool {
-	// fast path
-	if cr.Spec.IngestOnlyMode {
-		return true
-	}
-	return cr.Spec.isScrapeConfigUnmanaged()
 }
 
 // LastAppliedSpecAsPatch return last applied cluster spec as patch annotation

config/examples/vmagent-full.yaml

@@ -83,7 +83,7 @@ spec:
     repository: victoriametrics/vmagent
     tag: v1.46.0
     pullPolicy: IfNotPresent
-  vmAgentExternalLabelName: vmagent
+  externalLabelName: vmagent
   scrapeInterval: 30s
   externalLabels:
     key: value

docs/api.md

@@ -1757,6 +1757,7 @@ Appears in: [VMAgentSpec](#vmagentspec)
 | globalScrapeMetricRelabelConfigs<a href="#commonscrapeparams-globalscrapemetricrelabelconfigs" id="commonscrapeparams-globalscrapemetricrelabelconfigs">#</a><br/>_[RelabelConfig](#relabelconfig) array_ | _(Optional)_<br/>GlobalScrapeMetricRelabelConfigs is a global metric relabel configuration, which is applied to each scrape job. |
 | globalScrapeRelabelConfigs<a href="#commonscrapeparams-globalscraperelabelconfigs" id="commonscrapeparams-globalscraperelabelconfigs">#</a><br/>_[RelabelConfig](#relabelconfig) array_ | _(Optional)_<br/>GlobalScrapeRelabelConfigs is a global relabel configuration, which is applied to each samples of each scrape job during service discovery. |
 | ignoreNamespaceSelectors<a href="#commonscrapeparams-ignorenamespaceselectors" id="commonscrapeparams-ignorenamespaceselectors">#</a><br/>_boolean_ | _(Optional)_<br/>IgnoreNamespaceSelectors if set to true will ignore NamespaceSelector settings from<br />scrape objects, and they will only discover endpoints<br />within their current namespace. Defaults to false. |
+| ingestOnlyMode<a href="#commonscrapeparams-ingestonlymode" id="commonscrapeparams-ingestonlymode">#</a><br/>_boolean_ | _(Optional)_<br/>IngestOnlyMode switches vmagent into unmanaged mode<br />it disables any config generation for scraping<br />Currently it prevents vmagent from managing tls and auth options for remote write |
 | inlineScrapeConfig<a href="#commonscrapeparams-inlinescrapeconfig" id="commonscrapeparams-inlinescrapeconfig">#</a><br/>_string_ | _(Optional)_<br/>InlineScrapeConfig As scrape configs are appended, the user is responsible to make sure it<br />is valid. Note that using this feature may expose the possibility to<br />break upgrades of VMAgent. It is advised to review VMAgent release<br />notes to ensure that no incompatible scrape configs are going to break<br />VMAgent after the upgrade.<br />it should be defined as single yaml file.<br />inlineScrapeConfig: \|<br />    - job_name: "prometheus"<br />      static_configs:<br />      - targets: ["localhost:9090"] |
 | maxScrapeInterval<a href="#commonscrapeparams-maxscrapeinterval" id="commonscrapeparams-maxscrapeinterval">#</a><br/>_string_ | _(Required)_<br/>MaxScrapeInterval allows limiting maximum scrape interval for VMServiceScrape, VMPodScrape and other scrapes<br />If interval is higher than defined limit, `maxScrapeInterval` will be used. |
 | minScrapeInterval<a href="#commonscrapeparams-minscrapeinterval" id="commonscrapeparams-minscrapeinterval">#</a><br/>_string_ | _(Required)_<br/>MinScrapeInterval allows limiting minimal scrape interval for VMServiceScrape, VMPodScrape and other scrapes<br />If interval is lower than defined limit, `minScrapeInterval` will be used. |

docs/resources/vmagent.md

@@ -193,7 +193,7 @@ metadata:
 spec:
   # ...
   selectAllByDefault: true
-  vmAgentExternalLabelName: vmagent_ha
+  externalLabelName: vmagent_ha
   remoteWrite:
     - url: "http://vmsingle-example.default.svc:8428/api/v1/write"
   scrapeInterval: 30s
@@ -225,7 +225,7 @@ metadata:
 spec:
   # ...
   selectAllByDefault: true
-  vmAgentExternalLabelName: vmagent_ha
+  externalLabelName: vmagent_ha
   remoteWrite:
     - url: "http://vmsingle-example.default.svc:8428/api/v1/write"
   scrapeInterval: 30s
@@ -259,7 +259,7 @@ metadata:
 spec:
   # ...
   selectAllByDefault: true
-  vmAgentExternalLabelName: vmagent_ha
+  externalLabelName: vmagent_ha
   remoteWrite:
     - url: "http://vmsingle-example.default.svc:8428/api/v1/write"
   # Replication:
@@ -785,7 +785,7 @@ spec:
   scrapeTimeout: 10s
   externalLabels:
     cluster: my-cluster
-  vmAgentExternalLabelName: example
+  externalLabelName: example
   remoteWrite:
     - url: "http://vmsingle-example.default.svc:8428/api/v1/write"
   inlineRelabelConfig:

internal/controller/operator/factory/build/defaults.go

@@ -212,6 +212,9 @@ func addVMAgentDefaults(objI any) {
 	cv := config.ApplicationDefaults(c.VMAgentDefault)
 	addDefaultsToCommonParams(&cr.Spec.CommonDefaultableParams, cr.Spec.License, &cv)
 	addDefaultsToConfigReloader(&cr.Spec.CommonConfigReloaderParams, ptr.Deref(cr.Spec.UseDefaultResources, false))
+	if cr.Spec.IngestOnlyMode == nil {
+		cr.Spec.IngestOnlyMode = ptr.To(false)
+	}
 }
 
 func addVLAgentDefaults(objI any) {