Add query execution metrics (opensearch-project#799)

* Add query execution metrics Signed-off-by: Tomoyuki Morita <[email protected]> * Remove batch.processingTime metrics Signed-off-by: Tomoyuki Morita <[email protected]> --------- Signed-off-by: Tomoyuki Morita <[email protected]>
YANG-DB · Oct 23, 2024 · 7546f58 · 7546f58
1 parent 0b6da30
commit 7546f58
Show file tree

Hide file tree

Showing 4 changed files with 55 additions and 5 deletions.
diff --git a/flint-core/src/main/java/org/opensearch/flint/core/metrics/MetricConstants.java b/flint-core/src/main/java/org/opensearch/flint/core/metrics/MetricConstants.java
@@ -112,6 +112,11 @@ public final class MetricConstants {
      */
     public static final String STREAMING_HEARTBEAT_FAILED_METRIC = "streaming.heartbeat.failed.count";
 
+    /**
+     * Metric for tracking the latency of query execution (start to complete query execution) excluding result write.
+     */
+    public static final String QUERY_EXECUTION_TIME_METRIC = "query.execution.processingTime";
+
     private MetricConstants() {
         // Private constructor to prevent instantiation
     }

diff --git a/flint-core/src/main/java/org/opensearch/flint/core/metrics/MetricsUtil.java b/flint-core/src/main/java/org/opensearch/flint/core/metrics/MetricsUtil.java
@@ -107,6 +107,10 @@ public static Long stopTimer(Timer.Context context) {
         return context != null ? context.stop() : null;
     }
 
+    public static Timer getTimer(String metricName, boolean isIndexMetric) {
+        return getOrCreateTimer(metricName, isIndexMetric);
+    }
+
     /**
      * Registers a gauge metric with the provided name and value.
      *

diff --git a/flint-core/src/test/java/org/opensearch/flint/core/metrics/MetricsUtilTest.java b/flint-core/src/test/java/org/opensearch/flint/core/metrics/MetricsUtilTest.java
@@ -3,6 +3,8 @@
 import com.codahale.metrics.Counter;
 import com.codahale.metrics.Gauge;
 import com.codahale.metrics.Timer;
+import java.time.Duration;
+import java.time.temporal.TemporalUnit;
 import org.apache.spark.SparkEnv;
 import org.apache.spark.metrics.source.FlintMetricSource;
 import org.apache.spark.metrics.source.FlintIndexMetricSource;
@@ -101,6 +103,34 @@ private void testStartStopTimerHelper(boolean isIndexMetric) {
         }
     }
 
+    @Test
+    public void testGetTimer() {
+        try (MockedStatic<SparkEnv> sparkEnvMock = mockStatic(SparkEnv.class)) {
+            // Mock SparkEnv
+            SparkEnv sparkEnv = mock(SparkEnv.class, RETURNS_DEEP_STUBS);
+            sparkEnvMock.when(SparkEnv::get).thenReturn(sparkEnv);
+
+            // Mock appropriate MetricSource
+            String sourceName = FlintMetricSource.FLINT_INDEX_METRIC_SOURCE_NAME();
+            Source metricSource = Mockito.spy(new FlintIndexMetricSource());
+            when(sparkEnv.metricsSystem().getSourcesByName(sourceName).head()).thenReturn(
+                metricSource);
+
+            // Test the methods
+            String testMetric = "testPrefix.processingTime";
+            long duration = 500;
+            MetricsUtil.getTimer(testMetric, true).update(duration, TimeUnit.MILLISECONDS);
+
+            // Verify interactions
+            verify(sparkEnv.metricsSystem(), times(0)).registerSource(any());
+            verify(metricSource, times(1)).metricRegistry();
+            Timer timer = metricSource.metricRegistry().getTimers().get(testMetric);
+            Assertions.assertNotNull(timer);
+            Assertions.assertEquals(1L, timer.getCount());
+            assertEquals(Duration.ofMillis(duration).getNano(), timer.getSnapshot().getMean(), 0.1);
+        }
+    }
+
     @Test
     public void testRegisterGauge() {
         testRegisterGaugeHelper(false);
@@ -169,4 +199,4 @@ public void testDefaultBehavior() {
             Assertions.assertNotNull(flintMetricSource.metricRegistry().getGauges().get(testGaugeMetric));
         }
     }
-}
+}
diff --git a/spark-sql-application/src/main/scala/org/apache/spark/sql/JobOperator.scala b/spark-sql-application/src/main/scala/org/apache/spark/sql/JobOperator.scala
@@ -5,7 +5,7 @@
 
 package org.apache.spark.sql
 
-import java.util.concurrent.ThreadPoolExecutor
+import java.util.concurrent.{ThreadPoolExecutor, TimeUnit}
 import java.util.concurrent.atomic.AtomicInteger
 
 import scala.concurrent.{ExecutionContext, Future, TimeoutException}
@@ -14,7 +14,7 @@ import scala.util.{Failure, Success, Try}
 
 import org.opensearch.flint.common.model.FlintStatement
 import org.opensearch.flint.common.scheduler.model.LangType
-import org.opensearch.flint.core.metrics.MetricConstants
+import org.opensearch.flint.core.metrics.{MetricConstants, MetricsUtil}
 import org.opensearch.flint.core.metrics.MetricsUtil.incrementCounter
 import org.opensearch.flint.spark.FlintSpark
 
@@ -136,6 +136,8 @@ case class JobOperator(
             "",
             startTime))
     } finally {
+      emitQueryExecutionTimeMetric(startTime)
+
       try {
         dataToWrite.foreach(df => writeDataFrameToOpensearch(df, resultIndex, osClient))
       } catch {
@@ -148,11 +150,14 @@ case class JobOperator(
       statement.error = Some(error)
       statementExecutionManager.updateStatement(statement)
 
-      cleanUpResources(exceptionThrown, threadPool)
+      cleanUpResources(exceptionThrown, threadPool, startTime)
     }
   }
 
-  def cleanUpResources(exceptionThrown: Boolean, threadPool: ThreadPoolExecutor): Unit = {
+  def cleanUpResources(
+      exceptionThrown: Boolean,
+      threadPool: ThreadPoolExecutor,
+      startTime: Long): Unit = {
     val isStreaming = jobType.equalsIgnoreCase(FlintJobType.STREAMING)
     try {
       // Wait for streaming job complete if no error
@@ -195,6 +200,12 @@ case class JobOperator(
     }
   }
 
+  private def emitQueryExecutionTimeMetric(startTime: Long): Unit = {
+    MetricsUtil
+      .getTimer(MetricConstants.QUERY_EXECUTION_TIME_METRIC, false)
+      .update(System.currentTimeMillis() - startTime, TimeUnit.MILLISECONDS);
+  }
+
   def stop(): Unit = {
     Try {
       logInfo("Stopping Spark session")