[processor/resourcedetectionprocessor] Support Azure Tags

* Mirrors EC2 regex logic

Author: rimitchell

internal/metadataproviders/azure/metadata.go

@@ -37,15 +37,21 @@ func NewProvider() Provider {
 	}
 }
 
+type ComputeTagsListMetadata struct {
+	Name  string `json:"name"`
+	Value string `json:"value"`
+}
+
 // ComputeMetadata is the Azure IMDS compute metadata response format
 type ComputeMetadata struct {
-	Location          string `json:"location"`
-	Name              string `json:"name"`
-	VMID              string `json:"vmID"`
-	VMSize            string `json:"vmSize"`
-	SubscriptionID    string `json:"subscriptionID"`
-	ResourceGroupName string `json:"resourceGroupName"`
-	VMScaleSetName    string `json:"vmScaleSetName"`
+	Location          string                    `json:"location"`
+	Name              string                    `json:"name"`
+	VMID              string                    `json:"vmID"`
+	VMSize            string                    `json:"vmSize"`
+	SubscriptionID    string                    `json:"subscriptionID"`
+	ResourceGroupName string                    `json:"resourceGroupName"`
+	VMScaleSetName    string                    `json:"vmScaleSetName"`
+	TagsList          []ComputeTagsListMetadata `json:"tagsList"`
 }
 
 // Metadata queries a given endpoint and parses the output to the Azure IMDS format

internal/metadataproviders/azure/metadata_test.go

@@ -57,6 +57,12 @@ func TestQueryEndpointCorrect(t *testing.T) {
 		VMSize:            "vmSize",
 		SubscriptionID:    "subscriptionID",
 		ResourceGroupName: "resourceGroup",
+		TagsList: []ComputeTagsListMetadata{
+			{
+				Name:  "tag1",
+				Value: "value1",
+			},
+		},
 	}
 	marshalledMetadata, err := json.Marshal(sentMetadata)
 	require.NoError(t, err)

processor/resourcedetectionprocessor/README.md

@@ -397,6 +397,26 @@ processors:
     override: false
 ```
 
+It also can optionally gather tags from the azure instance that the collector is running on.
+
+Azure custom configuration example:
+
+```yaml
+processors:
+  resourcedetection/azure:
+    detectors: ["azure"]
+    azure:
+      # A list of regex's to match tag keys to add as resource attributes can be specified
+      tags:
+        - ^tag1$
+        - ^tag2$
+        - ^label.*$
+```
+
+Matched tags are added as:
+
+    * azure.tags.<tag name>
+
 ### Azure AKS
 
   * cloud.provider ("azure")

processor/resourcedetectionprocessor/internal/azure/azure.go

@@ -5,6 +5,7 @@ package azure // import "github.com/open-telemetry/opentelemetry-collector-contr
 
 import (
 	"context"
+	"regexp"
 
 	"go.opentelemetry.io/collector/pdata/pcommon"
 	"go.opentelemetry.io/collector/processor"
@@ -18,25 +19,34 @@ import (
 
 const (
 	// TypeStr is type of detector.
-	TypeStr = "azure"
+	TypeStr   = "azure"
+	tagPrefix = "azure.tag."
 )
 
 var _ internal.Detector = (*Detector)(nil)
 
 // Detector is an Azure metadata detector
 type Detector struct {
-	provider azure.Provider
-	logger   *zap.Logger
-	rb       *metadata.ResourceBuilder
+	provider      azure.Provider
+	tagKeyRegexes []*regexp.Regexp
+	logger        *zap.Logger
+	rb            *metadata.ResourceBuilder
 }
 
 // NewDetector creates a new Azure metadata detector
 func NewDetector(p processor.CreateSettings, dcfg internal.DetectorConfig) (internal.Detector, error) {
 	cfg := dcfg.(Config)
+
+	tagKeyRegexes, err := compileRegexes(cfg)
+	if err != nil {
+		return nil, err
+	}
+
 	return &Detector{
-		provider: azure.NewProvider(),
-		logger:   p.Logger,
-		rb:       metadata.NewResourceBuilder(cfg.ResourceAttributes),
+		provider:      azure.NewProvider(),
+		tagKeyRegexes: tagKeyRegexes,
+		logger:        p.Logger,
+		rb:            metadata.NewResourceBuilder(cfg.ResourceAttributes),
 	}, nil
 }
 
@@ -62,6 +72,47 @@ func (d *Detector) Detect(ctx context.Context) (resource pcommon.Resource, schem
 	d.rb.SetAzureVMSize(compute.VMSize)
 	d.rb.SetAzureVMScalesetName(compute.VMScaleSetName)
 	d.rb.SetAzureResourcegroupName(compute.ResourceGroupName)
+	res := d.rb.Emit()
+
+	if len(d.tagKeyRegexes) != 0 {
+		tags := matchAzureTags(compute.TagsList, d.tagKeyRegexes)
+		for key, val := range tags {
+			res.Attributes().PutStr(tagPrefix+key, val)
+		}
+	}
+
+	return res, conventions.SchemaURL, nil
+}
 
-	return d.rb.Emit(), conventions.SchemaURL, nil
+func matchAzureTags(azureTags []azure.ComputeTagsListMetadata, tagKeyRegexes []*regexp.Regexp) map[string]string {
+	tags := make(map[string]string)
+	for _, tag := range azureTags {
+		matched := regexArrayMatch(tagKeyRegexes, tag.Name)
+		if matched {
+			tags[tag.Name] = tag.Value
+		}
+	}
+	return tags
+}
+
+func compileRegexes(cfg Config) ([]*regexp.Regexp, error) {
+	tagRegexes := make([]*regexp.Regexp, len(cfg.Tags))
+	for i, elem := range cfg.Tags {
+		regex, err := regexp.Compile(elem)
+		if err != nil {
+			return nil, err
+		}
+		tagRegexes[i] = regex
+	}
+	return tagRegexes, nil
+}
+
+func regexArrayMatch(arr []*regexp.Regexp, val string) bool {
+	for _, elem := range arr {
+		matched := elem.MatchString(val)
+		if matched {
+			return true
+		}
+	}
+	return false
 }

processor/resourcedetectionprocessor/internal/azure/azure_test.go

@@ -6,6 +6,7 @@ package azure
 import (
 	"context"
 	"fmt"
+	"regexp"
 	"testing"
 
 	"github.com/stretchr/testify/assert"
@@ -36,9 +37,25 @@ func TestDetectAzureAvailable(t *testing.T) {
 		SubscriptionID:    "subscriptionID",
 		ResourceGroupName: "resourceGroup",
 		VMScaleSetName:    "myScaleset",
+		TagsList: []azure.ComputeTagsListMetadata{
+			{
+				Name:  "tag1key",
+				Value: "value1",
+			},
+			{
+				Name:  "tag2key",
+				Value: "value2",
+			},
+		},
 	}, nil)
 
-	detector := &Detector{provider: mp, rb: metadata.NewResourceBuilder(metadata.DefaultResourceAttributesConfig())}
+	detector := &Detector{
+		provider: mp,
+		tagKeyRegexes: []*regexp.Regexp{
+			regexp.MustCompile("^tag1key$"),
+		},
+		rb: metadata.NewResourceBuilder(metadata.DefaultResourceAttributesConfig()),
+	}
 	res, schemaURL, err := detector.Detect(context.Background())
 	require.NoError(t, err)
 	assert.Equal(t, conventions.SchemaURL, schemaURL)
@@ -55,9 +72,15 @@ func TestDetectAzureAvailable(t *testing.T) {
 		"azure.vm.size":                     "vmSize",
 		"azure.resourcegroup.name":          "resourceGroup",
 		"azure.vm.scaleset.name":            "myScaleset",
+		"azure.tag.tag1key":                 "value1",
+	}
+
+	notExpected := map[string]any{
+		"azure.tag.tag2key": "value2",
 	}
 
 	assert.Equal(t, expected, res.Attributes().AsRaw())
+	assert.NotEqual(t, notExpected, res.Attributes().AsRaw())
 }
 
 func TestDetectError(t *testing.T) {

processor/resourcedetectionprocessor/internal/azure/config.go

@@ -8,11 +8,15 @@ import (
 )
 
 type Config struct {
+	// Tags is a list of regex's to match ec2 instance tag keys that users want
+	// to add as resource attributes to processed data
+	Tags               []string                          `mapstructure:"tags"`
 	ResourceAttributes metadata.ResourceAttributesConfig `mapstructure:"resource_attributes"`
 }
 
 func CreateDefaultConfig() Config {
 	return Config{
+		Tags:               []string{},
 		ResourceAttributes: metadata.DefaultResourceAttributesConfig(),
 	}
 }