[elasticsearchexporter]: Add dynamic document id support for logs (#37065)

#### Description

This PR adds a new config option `logs_dynamic_id` that when set to true
reads the `elasticsearch.document_id` attribute from each log record and
uses it as the final document id in Elasticsearch. This is only
implemented for logs but I can open subsequent PRs supporting metrics
and traces akin to the `*_dynamic_index` options.

Fixes #36882

#### Testing

Added tests to verify that the document ID attribute can be read from
the log record and that the _id is properly forwarded to Elasticsearch.
Also asserted that when there is no doc id attribute the current
behavior is retained.

#### Documentation

Updated the readme to mention the new `logs_dynamic_id` config option.

---------

Co-authored-by: Carson Ip <[email protected]>
Co-authored-by: Christos Markou <[email protected]>

Author: 3 people

.chloggen/elasticsearchexporter_logs_dynamic_id.yaml

@@ -0,0 +1,27 @@
+# Use this changelog template to create an entry for release notes.
+
+# One of 'breaking', 'deprecation', 'new_component', 'enhancement', 'bug_fix'
+change_type: enhancement
+
+# The name of the component, or a single word describing the area of concern, (e.g. filelogreceiver)
+component: elasticsearchexporter
+
+# A brief description of the change.  Surround your text with quotes ("") if it needs to start with a backtick (`).
+note: Add config `logs_dynamic_id` to dynamically set the document ID of log records using log record attribute `elasticsearch.document_id`
+
+# Mandatory: One or more tracking issues related to the change. You can use the PR number here if no issue exists.
+issues: [36882]
+
+# (Optional) One or more lines of additional information to render under the primary note.
+# These lines will be padded with 2 spaces and then inserted directly into the document.
+# Use pipe (|) for multiline entries.
+subtext:
+
+# If your change doesn't affect end users or the exported elements of any package,
+# you should instead start your pull request title with [chore] or use the "Skip Changelog" label.
+# Optional: The change log or logs in which this entry should be included.
+# e.g. '[user]' or '[user, api]'
+# Include 'user' if the change is relevant to end users.
+# Include 'api' if there is a change to a library API.
+# Default: '[user]'
+change_logs: [user]

exporter/elasticsearchexporter/README.md

@@ -142,6 +142,9 @@ This can be customised through the following settings:
   - `prefix_separator`(default=`-`): Set a separator between logstash_prefix and date.
   - `date_format`(default=`%Y.%m.%d`): Time format (based on strftime) to generate the second part of the Index name.
 
+- `logs_dynamic_id` (optional): Dynamically determines the document ID to be used in Elasticsearch based on a log record attribute.
+  - `enabled`(default=false): Enable/Disable dynamic ID for log records. If `elasticsearch.document_id` exists and is not an empty string in the log record attributes, it will be used as the document ID. Otherwise, the document ID will be generated by Elasticsearch. The attribute `elasticsearch.document_id` is removed from the final document.
+
 ### Elasticsearch document mapping
 
 The Elasticsearch exporter supports several document schemas and preprocessing

exporter/elasticsearchexporter/bulkindexer.go

@@ -31,7 +31,7 @@ type bulkIndexer interface {
 
 type bulkIndexerSession interface {
 	// Add adds a document to the bulk indexing session.
-	Add(ctx context.Context, index string, document io.WriterTo, dynamicTemplates map[string]string) error
+	Add(ctx context.Context, index string, docID string, document io.WriterTo, dynamicTemplates map[string]string) error
 
 	// End must be called on the session object once it is no longer
 	// needed, in order to release any associated resources.
@@ -126,8 +126,9 @@ type syncBulkIndexerSession struct {
 }
 
 // Add adds an item to the sync bulk indexer session.
-func (s *syncBulkIndexerSession) Add(ctx context.Context, index string, document io.WriterTo, dynamicTemplates map[string]string) error {
-	err := s.bi.Add(docappender.BulkIndexerItem{Index: index, Body: document, DynamicTemplates: dynamicTemplates})
+func (s *syncBulkIndexerSession) Add(ctx context.Context, index string, docID string, document io.WriterTo, dynamicTemplates map[string]string) error {
+	doc := docappender.BulkIndexerItem{Index: index, Body: document, DocumentID: docID, DynamicTemplates: dynamicTemplates}
+	err := s.bi.Add(doc)
 	if err != nil {
 		return err
 	}
@@ -248,10 +249,11 @@ func (a *asyncBulkIndexer) Close(ctx context.Context) error {
 // Add adds an item to the async bulk indexer session.
 //
 // Adding an item after a call to Close() will panic.
-func (s asyncBulkIndexerSession) Add(ctx context.Context, index string, document io.WriterTo, dynamicTemplates map[string]string) error {
+func (s asyncBulkIndexerSession) Add(ctx context.Context, index string, docID string, document io.WriterTo, dynamicTemplates map[string]string) error {
 	item := docappender.BulkIndexerItem{
 		Index:            index,
 		Body:             document,
+		DocumentID:       docID,
 		DynamicTemplates: dynamicTemplates,
 	}
 	select {

exporter/elasticsearchexporter/bulkindexer_test.go

@@ -102,7 +102,7 @@ func TestAsyncBulkIndexer_flush(t *testing.T) {
 			session, err := bulkIndexer.StartSession(context.Background())
 			require.NoError(t, err)
 
-			assert.NoError(t, session.Add(context.Background(), "foo", strings.NewReader(`{"foo": "bar"}`), nil))
+			assert.NoError(t, session.Add(context.Background(), "foo", "", strings.NewReader(`{"foo": "bar"}`), nil))
 			// should flush
 			time.Sleep(100 * time.Millisecond)
 			assert.Equal(t, int64(1), bulkIndexer.stats.docsIndexed.Load())
@@ -229,7 +229,7 @@ func TestAsyncBulkIndexer_flush_error(t *testing.T) {
 			session, err := bulkIndexer.StartSession(context.Background())
 			require.NoError(t, err)
 
-			assert.NoError(t, session.Add(context.Background(), "foo", strings.NewReader(`{"foo": "bar"}`), nil))
+			assert.NoError(t, session.Add(context.Background(), "foo", "", strings.NewReader(`{"foo": "bar"}`), nil))
 			// should flush
 			time.Sleep(100 * time.Millisecond)
 			assert.Equal(t, int64(0), bulkIndexer.stats.docsIndexed.Load())
@@ -312,7 +312,7 @@ func runBulkIndexerOnce(t *testing.T, config *Config, client *elasticsearch.Clie
 	session, err := bulkIndexer.StartSession(context.Background())
 	require.NoError(t, err)
 
-	assert.NoError(t, session.Add(context.Background(), "foo", strings.NewReader(`{"foo": "bar"}`), nil))
+	assert.NoError(t, session.Add(context.Background(), "foo", "", strings.NewReader(`{"foo": "bar"}`), nil))
 	assert.NoError(t, bulkIndexer.Close(context.Background()))
 
 	return bulkIndexer
@@ -338,7 +338,7 @@ func TestSyncBulkIndexer_flushBytes(t *testing.T) {
 	session, err := bi.StartSession(context.Background())
 	require.NoError(t, err)
 
-	assert.NoError(t, session.Add(context.Background(), "foo", strings.NewReader(`{"foo": "bar"}`), nil))
+	assert.NoError(t, session.Add(context.Background(), "foo", "", strings.NewReader(`{"foo": "bar"}`), nil))
 	assert.Equal(t, int64(1), reqCnt.Load()) // flush due to flush::bytes
 	assert.NoError(t, bi.Close(context.Background()))
 }

exporter/elasticsearchexporter/config.go

@@ -53,6 +53,9 @@ type Config struct {
 	// fall back to pure TracesIndex, if 'elasticsearch.index.prefix' or 'elasticsearch.index.suffix' are not found in resource or attribute (prio: resource > attribute)
 	TracesDynamicIndex DynamicIndexSetting `mapstructure:"traces_dynamic_index"`
 
+	// LogsDynamicID configures whether log record attribute `elasticsearch.document_id` is set as the document ID in ES.
+	LogsDynamicID DynamicIDSettings `mapstructure:"logs_dynamic_id"`
+
 	// Pipeline configures the ingest node pipeline name that should be used to process the
 	// events.
 	//
@@ -112,6 +115,10 @@ type DynamicIndexSetting struct {
 	Enabled bool `mapstructure:"enabled"`
 }
 
+type DynamicIDSettings struct {
+	Enabled bool `mapstructure:"enabled"`
+}
+
 // AuthenticationSettings defines user authentication related settings.
 type AuthenticationSettings struct {
 	// User is used to configure HTTP Basic Authentication.

exporter/elasticsearchexporter/config_test.go

@@ -73,6 +73,9 @@ func TestConfig(t *testing.T) {
 				TracesDynamicIndex: DynamicIndexSetting{
 					Enabled: false,
 				},
+				LogsDynamicID: DynamicIDSettings{
+					Enabled: false,
+				},
 				Pipeline: "mypipeline",
 				ClientConfig: withDefaultHTTPClientConfig(func(cfg *confighttp.ClientConfig) {
 					cfg.Timeout = 2 * time.Minute
@@ -144,6 +147,9 @@ func TestConfig(t *testing.T) {
 				TracesDynamicIndex: DynamicIndexSetting{
 					Enabled: false,
 				},
+				LogsDynamicID: DynamicIDSettings{
+					Enabled: false,
+				},
 				Pipeline: "mypipeline",
 				ClientConfig: withDefaultHTTPClientConfig(func(cfg *confighttp.ClientConfig) {
 					cfg.Timeout = 2 * time.Minute
@@ -215,6 +221,9 @@ func TestConfig(t *testing.T) {
 				TracesDynamicIndex: DynamicIndexSetting{
 					Enabled: false,
 				},
+				LogsDynamicID: DynamicIDSettings{
+					Enabled: false,
+				},
 				Pipeline: "mypipeline",
 				ClientConfig: withDefaultHTTPClientConfig(func(cfg *confighttp.ClientConfig) {
 					cfg.Timeout = 2 * time.Minute

exporter/elasticsearchexporter/exporter.go

@@ -22,6 +22,11 @@ import (
 	"github.com/open-telemetry/opentelemetry-collector-contrib/exporter/elasticsearchexporter/internal/pool"
 )
 
+const (
+	// documentIDAttributeName is the attribute name used to specify the document ID.
+	documentIDAttributeName = "elasticsearch.document_id"
+)
+
 type elasticsearchExporter struct {
 	component.TelemetrySettings
 	userAgent string
@@ -176,13 +181,15 @@ func (e *elasticsearchExporter) pushLogRecord(
 	}
 
 	buf := e.bufferPool.NewPooledBuffer()
+	docID := e.extractDocumentIDAttribute(record.Attributes())
 	err := e.model.encodeLog(resource, resourceSchemaURL, record, scope, scopeSchemaURL, fIndex, buf.Buffer)
 	if err != nil {
 		buf.Recycle()
 		return fmt.Errorf("failed to encode log event: %w", err)
 	}
+
 	// not recycling after Add returns an error as we don't know if it's already recycled
-	return bulkIndexerSession.Add(ctx, fIndex.Index, buf, nil)
+	return bulkIndexerSession.Add(ctx, fIndex.Index, docID, buf, nil)
 }
 
 func (e *elasticsearchExporter) pushMetricsData(
@@ -299,7 +306,7 @@ func (e *elasticsearchExporter) pushMetricsData(
 						errs = append(errs, err)
 						continue
 					}
-					if err := session.Add(ctx, fIndex.Index, buf, dynamicTemplates); err != nil {
+					if err := session.Add(ctx, fIndex.Index, "", buf, dynamicTemplates); err != nil {
 						// not recycling after Add returns an error as we don't know if it's already recycled
 						if cerr := ctx.Err(); cerr != nil {
 							return cerr
@@ -422,7 +429,7 @@ func (e *elasticsearchExporter) pushTraceRecord(
 		return fmt.Errorf("failed to encode trace record: %w", err)
 	}
 	// not recycling after Add returns an error as we don't know if it's already recycled
-	return bulkIndexerSession.Add(ctx, fIndex.Index, buf, nil)
+	return bulkIndexerSession.Add(ctx, fIndex.Index, "", buf, nil)
 }
 
 func (e *elasticsearchExporter) pushSpanEvent(
@@ -454,5 +461,17 @@ func (e *elasticsearchExporter) pushSpanEvent(
 		return nil
 	}
 	// not recycling after Add returns an error as we don't know if it's already recycled
-	return bulkIndexerSession.Add(ctx, fIndex.Index, buf, nil)
+	return bulkIndexerSession.Add(ctx, fIndex.Index, "", buf, nil)
+}
+
+func (e *elasticsearchExporter) extractDocumentIDAttribute(m pcommon.Map) string {
+	if !e.config.LogsDynamicID.Enabled {
+		return ""
+	}
+
+	v, ok := m.Get(documentIDAttributeName)
+	if !ok {
+		return ""
+	}
+	return v.AsString()
 }

exporter/elasticsearchexporter/exporter_test.go

@@ -736,6 +736,82 @@ func TestExporterLogs(t *testing.T) {
 		assert.JSONEq(t, `{"a":"a","a.b":"a.b"}`, gjson.GetBytes(doc, `resource.attributes`).Raw)
 	})
 
+	t.Run("publish logs with dynamic id", func(t *testing.T) {
+		t.Parallel()
+		exampleDocID := "abc123"
+		tableTests := []struct {
+			name          string
+			expectedDocID string // "" means the _id will not be set
+			recordAttrs   map[string]any
+		}{
+			{
+				name:          "missing document id attribute should not set _id",
+				expectedDocID: "",
+			},
+			{
+				name:          "empty document id attribute should not set _id",
+				expectedDocID: "",
+				recordAttrs: map[string]any{
+					documentIDAttributeName: "",
+				},
+			},
+			{
+				name:          "record attributes",
+				expectedDocID: exampleDocID,
+				recordAttrs: map[string]any{
+					documentIDAttributeName: exampleDocID,
+				},
+			},
+		}
+
+		cfgs := map[string]func(*Config){
+			"async": func(cfg *Config) {
+				batcherEnabled := false
+				cfg.Batcher.Enabled = &batcherEnabled
+			},
+			"sync": func(cfg *Config) {
+				batcherEnabled := true
+				cfg.Batcher.Enabled = &batcherEnabled
+				cfg.Batcher.FlushTimeout = 10 * time.Millisecond
+			},
+		}
+		for _, tt := range tableTests {
+			for cfgName, cfgFn := range cfgs {
+				t.Run(tt.name+"/"+cfgName, func(t *testing.T) {
+					t.Parallel()
+					rec := newBulkRecorder()
+					server := newESTestServer(t, func(docs []itemRequest) ([]itemResponse, error) {
+						rec.Record(docs)
+
+						if tt.expectedDocID == "" {
+							assert.NotContains(t, string(docs[0].Action), "_id", "expected _id to not be set")
+						} else {
+							assert.Equal(t, tt.expectedDocID, actionJSONToID(t, docs[0].Action), "expected _id to be set")
+						}
+
+						// Ensure the document id attribute is removed from the final document.
+						assert.NotContains(t, string(docs[0].Document), documentIDAttributeName, "expected document id attribute to be removed")
+						return itemsAllOK(docs)
+					})
+
+					exporter := newTestLogsExporter(t, server.URL, func(cfg *Config) {
+						cfg.Mapping.Mode = "otel"
+						cfg.LogsDynamicID.Enabled = true
+						cfgFn(cfg)
+					})
+					logs := newLogsWithAttributes(
+						tt.recordAttrs,
+						map[string]any{},
+						map[string]any{},
+					)
+					logs.ResourceLogs().At(0).ScopeLogs().At(0).LogRecords().At(0).Body().SetStr("hello world")
+					mustSendLogs(t, exporter, logs)
+
+					rec.WaitItems(1)
+				})
+			}
+		}
+	})
 	t.Run("otel mode attribute complex value", func(t *testing.T) {
 		rec := newBulkRecorder()
 		server := newESTestServer(t, func(docs []itemRequest) ([]itemResponse, error) {
@@ -1943,3 +2019,14 @@ func actionJSONToIndex(t *testing.T, actionJSON json.RawMessage) string {
 	require.NoError(t, err)
 	return action.Create.Index
 }
+
+func actionJSONToID(t *testing.T, actionJSON json.RawMessage) string {
+	action := struct {
+		Create struct {
+			ID string `json:"_id"`
+		} `json:"create"`
+	}{}
+	err := json.Unmarshal(actionJSON, &action)
+	require.NoError(t, err)
+	return action.Create.ID
+}

exporter/elasticsearchexporter/factory.go

@@ -62,6 +62,9 @@ func createDefaultConfig() component.Config {
 		TracesDynamicIndex: DynamicIndexSetting{
 			Enabled: false,
 		},
+		LogsDynamicID: DynamicIDSettings{
+			Enabled: false,
+		},
 		Retry: RetrySettings{
 			Enabled:         true,
 			MaxRetries:      0, // default is set in exporter code

exporter/elasticsearchexporter/pdata_serializer.go

@@ -298,7 +298,7 @@ func writeAttributes(v *json.Visitor, attributes pcommon.Map, stringifyMapValues
 	_ = v.OnObjectStart(-1, structform.AnyType)
 	attributes.Range(func(k string, val pcommon.Value) bool {
 		switch k {
-		case dataStreamType, dataStreamDataset, dataStreamNamespace, elasticsearch.MappingHintsAttrKey:
+		case dataStreamType, dataStreamDataset, dataStreamNamespace, elasticsearch.MappingHintsAttrKey, documentIDAttributeName:
 			return true
 		}
 		if isGeoAttribute(k, val) {

exporter/elasticsearchexporter/pdata_serializer_test.go

@@ -31,6 +31,7 @@ func TestSerializeLog(t *testing.T) {
 			record.Attributes().PutDouble("double", 42.0)
 			record.Attributes().PutInt("int", 42)
 			record.Attributes().PutEmptyBytes("bytes").Append(42)
+			record.Attributes().PutStr(documentIDAttributeName, "my_id")
 			_ = record.Attributes().PutEmptySlice("slice").FromRaw([]any{42, "foo"})
 			record.Attributes().PutEmptySlice("map_slice").AppendEmpty().SetEmptyMap().PutStr("foo.bar", "baz")
 			mapAttr := record.Attributes().PutEmptyMap("map")