Change type of ingest doc meta-data field 'TIMESTAMP' to `Date` (#22234)

With this commit we change the data type of the 'TIMESTAMP'
meta-data field from a formatted date string to a plain
`java.util.Date` instance. The main reason for this change is
that our benchmarks have indicated that this contributes
significantly to the time spent in the ingest pipeline.

The overhead in terms of indexing throughput of the ingest
pipeline is about 15% and breaks down roughly as follows:

* 5% overhead caused by the conversion from `XContent` -> `Map`
* 5% overhead caused by the timestamp formatting
* 5% overhead caused by the conversion `Map` -> `XContent`

Relates #22074

core/src/main/java/org/elasticsearch/ingest/IngestDocument.java

@@ -27,18 +27,14 @@ import org.elasticsearch.index.mapper.RoutingFieldMapper;
 import org.elasticsearch.index.mapper.SourceFieldMapper;
 import org.elasticsearch.index.mapper.TypeFieldMapper;
 
-import java.text.DateFormat;
-import java.text.SimpleDateFormat;
 import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.Base64;
 import java.util.Date;
 import java.util.HashMap;
 import java.util.List;
-import java.util.Locale;
 import java.util.Map;
 import java.util.Objects;
-import java.util.TimeZone;
 
 /**
  * Represents a single document being captured before indexing and holds the source and metadata (like id, type and index).
@@ -68,9 +64,7 @@ public final class IngestDocument {
         }
 
         this.ingestMetadata = new HashMap<>();
-        DateFormat df = new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ss.SSSZZ", Locale.ROOT);
+        this.ingestMetadata.put(TIMESTAMP, new Date());
-        df.setTimeZone(TimeZone.getTimeZone("UTC"));
-        this.ingestMetadata.put(TIMESTAMP, df.format(new Date()));
     }
 
     /**
@@ -595,6 +589,8 @@ public final class IngestDocument {
             value instanceof Long || value instanceof Float ||
             value instanceof Double || value instanceof Boolean) {
             return value;
+        } else if (value instanceof Date) {
+            return ((Date) value).clone();
         } else {
             throw new IllegalArgumentException("unexpected value type [" + value.getClass() + "]");
         }

core/src/test/java/org/elasticsearch/ingest/IngestDocumentTests.java

@@ -22,21 +22,17 @@ package org.elasticsearch.ingest;
 import org.elasticsearch.test.ESTestCase;
 import org.junit.Before;
 
-import java.text.DateFormat;
-import java.text.SimpleDateFormat;
 import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.Collections;
 import java.util.Date;
 import java.util.HashMap;
 import java.util.List;
-import java.util.Locale;
 import java.util.Map;
 
 import static org.elasticsearch.ingest.IngestDocumentMatcher.assertIngestDocument;
 import static org.hamcrest.Matchers.both;
 import static org.hamcrest.Matchers.containsString;
-import static org.hamcrest.Matchers.endsWith;
 import static org.hamcrest.Matchers.equalTo;
 import static org.hamcrest.Matchers.greaterThanOrEqualTo;
 import static org.hamcrest.Matchers.instanceOf;
@@ -48,13 +44,14 @@ import static org.hamcrest.Matchers.sameInstance;
 
 public class IngestDocumentTests extends ESTestCase {
 
+    private static final Date BOGUS_TIMESTAMP = new Date(0L);
     private IngestDocument ingestDocument;
 
     @Before
     public void setIngestDocument() {
         Map<String, Object> document = new HashMap<>();
         Map<String, Object> ingestMap = new HashMap<>();
-        ingestMap.put("timestamp", "bogus_timestamp");
+        ingestMap.put("timestamp", BOGUS_TIMESTAMP);
         document.put("_ingest", ingestMap);
         document.put("foo", "bar");
         document.put("int", 123);
@@ -86,9 +83,9 @@ public class IngestDocumentTests extends ESTestCase {
         assertThat(ingestDocument.getFieldValue("_index", String.class), equalTo("index"));
         assertThat(ingestDocument.getFieldValue("_type", String.class), equalTo("type"));
         assertThat(ingestDocument.getFieldValue("_id", String.class), equalTo("id"));
-        assertThat(ingestDocument.getFieldValue("_ingest.timestamp", String.class),
+        assertThat(ingestDocument.getFieldValue("_ingest.timestamp", Date.class),
-                both(notNullValue()).and(not(equalTo("bogus_timestamp"))));
+                both(notNullValue()).and(not(equalTo(BOGUS_TIMESTAMP))));
-        assertThat(ingestDocument.getFieldValue("_source._ingest.timestamp", String.class), equalTo("bogus_timestamp"));
+        assertThat(ingestDocument.getFieldValue("_source._ingest.timestamp", Date.class), equalTo(BOGUS_TIMESTAMP));
     }
 
     public void testGetSourceObject() {
@@ -972,11 +969,8 @@ public class IngestDocumentTests extends ESTestCase {
         long before = System.currentTimeMillis();
         IngestDocument ingestDocument = RandomDocumentPicks.randomIngestDocument(random());
         long after = System.currentTimeMillis();
-        String timestampString = (String) ingestDocument.getIngestMetadata().get("timestamp");
+        Date timestamp = (Date) ingestDocument.getIngestMetadata().get(IngestDocument.TIMESTAMP);
-        assertThat(timestampString, notNullValue());
+        assertThat(timestamp, notNullValue());
-        assertThat(timestampString, endsWith("+0000"));
-        DateFormat df = new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ss.SSSZZ", Locale.ROOT);
-        Date timestamp = df.parse(timestampString);
         assertThat(timestamp.getTime(), greaterThanOrEqualTo(before));
         assertThat(timestamp.getTime(), lessThanOrEqualTo(after));
     }

docs/reference/migration/migrate_6_0.asciidoc

@@ -35,6 +35,7 @@ way to reindex old indices is to use the `reindex` API.
 * <<breaking_60_plugins_changes>>
 * <<breaking_60_indices_changes>>
 * <<breaking_60_scripting_changes>>
+* <<breaking_60_ingest_changes>>
 
 include::migrate_6_0/cat.asciidoc[]
 
@@ -57,3 +58,5 @@ include::migrate_6_0/plugins.asciidoc[]
 include::migrate_6_0/indices.asciidoc[]
 
 include::migrate_6_0/scripting.asciidoc[]
+
+include::migrate_6_0/ingest.asciidoc[]

docs/reference/migration/migrate_6_0/ingest.asciidoc

@@ -0,0 +1,6 @@
+[[breaking_60_ingest_changes]]
+=== Ingest changes
+
+==== Timestamp meta-data field type has changed
+
+The type of the "timestamp" meta-data field has changed from `java.lang.String` to `java.util.Date`.