[extension/googlecloudlogentryencoding] Fix incorrect snake_case conversion for numeronyms (e.g. k8s)

.chloggen/fix_gcp_snake_case.yaml

@@ -0,0 +1,5 @@
+change_type: bug_fix
+component: extension/googlecloudlogentry_encoding
+note: Fix incorrect snake_case conversion for keys containing numbers (e.g., "k8s" becoming "k8_s") in Google Cloud log entries.
+issues: [46571]
+subtext: ""

extension/encoding/googlecloudlogentryencodingextension/internal/auditlog/parser.go

@@ -9,7 +9,6 @@ import (
 	"strings"
 
 	gojson "github.com/goccy/go-json"
-	"github.com/iancoleman/strcase"
 	"go.opentelemetry.io/collector/pdata/pcommon"
 	conventions "go.opentelemetry.io/otel/semconv/v1.38.0"
 
@@ -295,7 +294,7 @@ func handlePolicyViolationInfo(info *policyViolationInfo, attr pcommon.Map) {
 	if len(info.OrgPolicyViolationInfo.ResourceTags) > 0 {
 		tags := attr.PutEmptyMap(gcpAuditPolicyViolationResourceTags)
 		for name, value := range info.OrgPolicyViolationInfo.ResourceTags {
-			shared.PutStr(strcase.ToSnakeWithIgnore(name, "."), value, tags)
+			shared.PutStr(shared.ToSnakeCase(name, "."), value, tags)
 		}
 	}
 
@@ -365,7 +364,7 @@ func handleRequestMetadata(metadata *requestMetadata, attr pcommon.Map) error {
 		if len(metadata.DestinationAttributes.Labels) > 0 {
 			m := attr.PutEmptyMap(gcpAuditDestinationLabels)
 			for l, v := range metadata.DestinationAttributes.Labels {
-				shared.PutStr(strcase.ToSnakeWithIgnore(l, "."), v, m)
+				shared.PutStr(shared.ToSnakeCase(l, "."), v, m)
 			}
 		}
 	}

extension/encoding/googlecloudlogentryencodingextension/internal/shared/snakecase.go

@@ -0,0 +1,29 @@
+// Copyright The OpenTelemetry Authors
+// SPDX-License-Identifier: Apache-2.0
+
+package shared // import "github.com/open-telemetry/opentelemetry-collector-contrib/extension/encoding/googlecloudlogentryencodingextension/internal/shared"
+
+import (
+	"regexp"
+	"strings"
+
+	"github.com/iancoleman/strcase"
+)
+
+// strcase incorrectly treats digit-to-letter transitions as word boundaries,
+// inserting underscores (e.g. "k8s" -> "k8_s").
+// This regex reverses that.
+// See: https://github.com/open-telemetry/opentelemetry-collector-contrib/issues/46571
+var digitToLetter = regexp.MustCompile(`(\d)_([a-z])`)
+
+// ToSnakeCase converts a string to snake_case while preserving characters in
+// the ignore set, and fixes strcase's incorrect splitting at digit-to-letter
+// boundaries like "k8s" into "k8_s".
+func ToSnakeCase(s, ignore string) string {
+	result := strcase.ToSnakeWithIgnore(s, ignore)
+	// Fast path: if there are no digits, skip the regex entirely
+	if !strings.ContainsAny(result, "0123456789") {
+		return result
+	}
+	return digitToLetter.ReplaceAllString(result, "${1}${2}")
+}

extension/encoding/googlecloudlogentryencodingextension/log_entry.go

@@ -13,7 +13,6 @@ import (
 	"time"
 
 	gojson "github.com/goccy/go-json"
-	"github.com/iancoleman/strcase"
 	"go.opentelemetry.io/collector/pdata/pcommon"
 	"go.opentelemetry.io/collector/pdata/plog"
 	conventions "go.opentelemetry.io/otel/semconv/v1.38.0"
@@ -593,7 +592,7 @@ func handleLogEntryFields(resourceAttributes pcommon.Map, scopeLogs plog.ScopeLo
 	if log.Resource != nil {
 		resourceAttributes.PutStr(gcpResourceTypeField, log.Resource.Type)
 		for k, v := range log.Resource.Labels {
-			shared.PutStr(strcase.ToSnakeWithIgnore(fmt.Sprintf("gcp.label.%s", k), "."), v, resourceAttributes)
+			shared.PutStr(shared.ToSnakeCase(fmt.Sprintf("gcp.label.%s", k), "."), v, resourceAttributes)
 		}
 	}
 
@@ -624,7 +623,7 @@ func handleLogEntryFields(resourceAttributes pcommon.Map, scopeLogs plog.ScopeLo
 	}
 
 	for k, v := range log.Labels {
-		logRecord.Attributes().PutStr(strcase.ToSnakeWithIgnore(fmt.Sprintf("gcp.label.%v", k), "."), v)
+		logRecord.Attributes().PutStr(shared.ToSnakeCase(fmt.Sprintf("gcp.label.%v", k), "."), v)
 	}
 
 	handleOperationField(logRecord.Attributes(), log.Operation)

extension/encoding/googlecloudlogentryencodingextension/log_entry_test.go

@@ -8,6 +8,7 @@ import (
 	"testing"
 
 	gojson "github.com/goccy/go-json"
+	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"go.opentelemetry.io/collector/pdata/pcommon"
 	"go.opentelemetry.io/collector/pdata/plog"
@@ -17,6 +18,7 @@ import (
 	"github.com/open-telemetry/opentelemetry-collector-contrib/extension/encoding/googlecloudlogentryencodingextension/internal/constants"
 	"github.com/open-telemetry/opentelemetry-collector-contrib/extension/encoding/googlecloudlogentryencodingextension/internal/passthroughnlb"
 	"github.com/open-telemetry/opentelemetry-collector-contrib/extension/encoding/googlecloudlogentryencodingextension/internal/proxynlb"
+	"github.com/open-telemetry/opentelemetry-collector-contrib/extension/encoding/googlecloudlogentryencodingextension/internal/shared"
 	"github.com/open-telemetry/opentelemetry-collector-contrib/extension/encoding/googlecloudlogentryencodingextension/internal/vpcflowlog"
 	"github.com/open-telemetry/opentelemetry-collector-contrib/pkg/golden"
 	"github.com/open-telemetry/opentelemetry-collector-contrib/pkg/pdatatest/plogtest"
@@ -595,3 +597,43 @@ func TestHandleProtoPayload(t *testing.T) {
 		})
 	}
 }
+
+func TestToSnakeCase(t *testing.T) {
+	tests := []struct {
+		name     string
+		input    string
+		ignore   string
+		expected string
+	}{
+		{
+			name:     "k8s label key preserved",
+			input:    "labels.authorization.k8s.io/decision",
+			ignore:   ".",
+			expected: "labels.authorization.k8s.io/decision",
+		},
+		{
+			name:     "already snake_case unchanged",
+			input:    "already_snake",
+			ignore:   ".",
+			expected: "already_snake",
+		},
+		{
+			name:     "simple camelCase converted",
+			input:    "simpleLabel",
+			ignore:   ".",
+			expected: "simple_label",
+		},
+		{
+			name:     "h2c in lowercase preserved",
+			input:    "proxy.h2c.enabled",
+			ignore:   ".",
+			expected: "proxy.h2c.enabled",
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result := shared.ToSnakeCase(tt.input, tt.ignore)
+			assert.Equal(t, tt.expected, result)
+		})
+	}
+}

extension/encoding/googlecloudlogentryencodingextension/testdata/auditlog/activity_expected.yaml

@@ -61,10 +61,10 @@ resourceLogs:
               - key: user_agent.original
                 value:
                   stringValue: gcp-controller-manager/v0.0.0 (linux/amd64) kubernetes/$Format/leader-election
-              - key: gcp.label.authorization.k8_s.io/reason
+              - key: gcp.label.authorization.k8s.io/reason
                 value:
                   stringValue: 'RBAC: allowed by ClusterRoleBinding "system:gcp-controller-manager" of ClusterRole "system:gcp-controller-manager" to User "system:gcp-controller-manager"'
-              - key: gcp.label.authorization.k8_s.io/decision
+              - key: gcp.label.authorization.k8s.io/decision
                 value:
                   stringValue: allow
               - key: gcp.operation.id