[#3152] feat(core): using time sliding window to record timer metrics (…

…#3341)

### What changes were proposed in this pull request?

using time sliding window to record timer metrics

### Why are the changes needed?

Fix: #3152 

### Does this PR introduce _any_ user-facing change?
no

### How was this patch tested?
test in local env to check whether Pxx comes to 0 after a while

common/src/main/java/com/datastrato/gravitino/config/ConfigConstants.java

@@ -32,4 +32,6 @@ private ConfigConstants() {}
   public static final String VERSION_0_4_0 = "0.4.0";
   /** The version number for the 0.5.0 release. */
   public static final String VERSION_0_5_0 = "0.5.0";
+  /** The version number for the 0.5.1 release. */
+  public static final String VERSION_0_5_1 = "0.5.1";
 }

core/src/main/java/com/datastrato/gravitino/Configs.java

@@ -276,4 +276,12 @@ private Configs() {}
           .version(ConfigConstants.VERSION_0_5_0)
           .longConf()
           .createWithDefault(60 * 60 * 1000L);
+
+  public static final int DEFAULT_METRICS_TIME_SLIDING_WINDOW_SECONDS = 60;
+  public static final ConfigEntry<Integer> METRICS_TIME_SLIDING_WINDOW_SECONDS =
+      new ConfigBuilder("gravitino.metrics.timeSlidingWindowSecs")
+          .doc("The seconds of Gravitino metrics time sliding window")
+          .version(ConfigConstants.VERSION_0_5_1)
+          .intConf()
+          .createWithDefault(DEFAULT_METRICS_TIME_SLIDING_WINDOW_SECONDS);
 }

core/src/main/java/com/datastrato/gravitino/metrics/source/MetricsSource.java

@@ -9,7 +9,12 @@
 import com.codahale.metrics.Gauge;
 import com.codahale.metrics.Histogram;
 import com.codahale.metrics.MetricRegistry;
+import com.codahale.metrics.SlidingTimeWindowArrayReservoir;
 import com.codahale.metrics.Timer;
+import com.datastrato.gravitino.Config;
+import com.datastrato.gravitino.Configs;
+import com.datastrato.gravitino.GravitinoEnv;
+import java.util.concurrent.TimeUnit;
 
 /**
  * MetricsSource provides utilities to collect specified kind metrics, all metrics must create with
@@ -24,10 +29,19 @@ public abstract class MetricsSource {
   public static final String JVM_METRIC_NAME = "jvm";
   private final MetricRegistry metricRegistry;
   private final String metricsSourceName;
+  private final int timeSlidingWindowSeconds;
 
   protected MetricsSource(String name) {
     this.metricsSourceName = name;
     metricRegistry = new MetricRegistry();
+    Config config = GravitinoEnv.getInstance().config();
+    if (config != null) {
+      this.timeSlidingWindowSeconds =
+          config.get(Configs.METRICS_TIME_SLIDING_WINDOW_SECONDS).intValue();
+    } else {
+      // Couldn't get config when testing
+      this.timeSlidingWindowSeconds = Configs.DEFAULT_METRICS_TIME_SLIDING_WINDOW_SECONDS;
+    }
   }
 
   /**
@@ -75,7 +89,12 @@ public Counter getCounter(String name) {
    * @return a new or pre-existing Histogram
    */
   public Histogram getHistogram(String name) {
-    return this.metricRegistry.histogram(name);
+    return this.metricRegistry.histogram(
+        name,
+        () ->
+            new Histogram(
+                new SlidingTimeWindowArrayReservoir(
+                    getTimeSlidingWindowSeconds(), TimeUnit.SECONDS)));
   }
 
   /**
@@ -85,6 +104,15 @@ public Histogram getHistogram(String name) {
    * @return a new or pre-existing Timer
    */
   public Timer getTimer(String name) {
-    return this.metricRegistry.timer(name);
+    return this.metricRegistry.timer(
+        name,
+        () ->
+            new Timer(
+                new SlidingTimeWindowArrayReservoir(
+                    getTimeSlidingWindowSeconds(), TimeUnit.SECONDS)));
+  }
+
+  protected int getTimeSlidingWindowSeconds() {
+    return timeSlidingWindowSeconds;
   }
 }

core/src/test/java/com/datastrato/gravitino/metrics/TestMetricsSystem.java

@@ -10,11 +10,7 @@
 import org.junit.jupiter.api.Test;
 
 public class TestMetricsSystem {
-  MetricsSystem metricsSystem = new MetricsSystem();
-
-  private long getCounterValue(String metricsSourceName, String name) {
-    return metricsSystem.getMetricRegistry().counter(metricsSourceName + "." + name).getCount();
-  }
+  private MetricsSystem metricsSystem = new MetricsSystem();
 
   @Test
   void testRegisterMetricsSource() {
@@ -53,4 +49,8 @@ void testRegisterMetricsSource() {
             .getCounters()
             .containsKey(metricsSource.getMetricsSourceName() + "a.b"));
   }
+
+  private long getCounterValue(String metricsSourceName, String name) {
+    return metricsSystem.getMetricRegistry().counter(metricsSourceName + "." + name).getCount();
+  }
 }

core/src/test/java/com/datastrato/gravitino/metrics/source/TestMetricsSource.java

@@ -28,7 +28,7 @@ public TestMetricsSource() {
 
   @BeforeAll
   void init() {
-    metricsSystem = new MetricsSystem();
+    this.metricsSystem = new MetricsSystem();
     metricsSystem.register(this);
   }
 

docs/gravitino-server-config.md

@@ -131,6 +131,12 @@ For more details, please refer to the definition of the plugin.
 
 Refer to [security](security.md) for HTTPS and authentication configurations.
 
+### Metrics configuration
+
+| Property name                             | Description                                          | Default value | Required | Since Version |
+|-------------------------------------------|------------------------------------------------------|---------------|----------|---------------|
+| `gravitino.metrics.timeSlidingWindowSecs` | The seconds of Gravitino metrics time sliding window | 60            | No       | 0.5.1         |
+
 ## Gravitino catalog properties configuration
 
 There are three types of catalog properties:

server-common/src/main/java/com/datastrato/gravitino/server/web/HttpServerMetricsSource.java

@@ -5,15 +5,25 @@
 
 package com.datastrato.gravitino.server.web;
 
+import com.codahale.metrics.Clock;
+import com.codahale.metrics.SlidingTimeWindowArrayReservoir;
 import com.codahale.metrics.jersey2.InstrumentedResourceMethodApplicationListener;
 import com.datastrato.gravitino.metrics.MetricNames;
 import com.datastrato.gravitino.metrics.source.MetricsSource;
+import java.util.concurrent.TimeUnit;
 import org.glassfish.jersey.server.ResourceConfig;
 
 public class HttpServerMetricsSource extends MetricsSource {
   public HttpServerMetricsSource(String name, ResourceConfig resourceConfig, JettyServer server) {
     super(name);
-    resourceConfig.register(new InstrumentedResourceMethodApplicationListener(getMetricRegistry()));
+    resourceConfig.register(
+        new InstrumentedResourceMethodApplicationListener(
+            getMetricRegistry(),
+            Clock.defaultClock(),
+            false,
+            () ->
+                new SlidingTimeWindowArrayReservoir(
+                    getTimeSlidingWindowSeconds(), TimeUnit.SECONDS)));
     registerGauge(
         MetricNames.SERVER_IDLE_THREAD_NUM, () -> server.getThreadPool().getIdleThreads());
   }