Added optional tags/dimensions for PartitionId/ReplicaId as addition/… (

#35164) * Added optional tags/dimensions for PartitionId/ReplicaId as addition/alternative to ServiceAddress for rntbd/request-level metrics * Update CHANGELOG.md * Update Metrics.md * Update ClientTelemetryMetrics.java
Azure · May 30, 2023 · eb4b160 · eb4b160
1 parent 1a310ce
commit eb4b160
Show file tree

Hide file tree

Showing 6 changed files with 150 additions and 57 deletions.
diff --git a/sdk/cosmos/azure-cosmos-tests/src/test/java/com/azure/cosmos/ClientMetricsTest.java b/sdk/cosmos/azure-cosmos-tests/src/test/java/com/azure/cosmos/ClientMetricsTest.java
@@ -98,15 +98,21 @@ public void beforeTest(CosmosMetricCategory... metricCategories) {
         beforeTest(null, metricCategories);
     }
 
-    public void beforeTest(CosmosDiagnosticsThresholds thresholds, CosmosMetricCategory... metricCategories) {
+    public void beforeTest(
+        CosmosDiagnosticsThresholds thresholds,
+        CosmosMetricCategory... metricCategories) {
         assertThat(this.client).isNull();
         assertThat(this.meterRegistry).isNull();
 
         this.meterRegistry = ConsoleLoggingRegistryFactory.create(1);
 
         this.inputMetricsOptions = new CosmosMicrometerMetricsOptions()
             .meterRegistry(this.meterRegistry)
-            .setMetricCategories(metricCategories);
+            .setMetricCategories(metricCategories)
+            .configureDefaultTagNames(
+                CosmosMetricTagName.DEFAULT,
+                CosmosMetricTagName.PARTITION_ID,
+                CosmosMetricTagName.REPLICA_ID);
 
         this.inputClientTelemetryConfig = new CosmosClientTelemetryConfig()
             .metricsOptions(this.inputMetricsOptions);
@@ -234,13 +240,31 @@ public void createItem() throws Exception {
                 this.assertMetrics("cosmos.client.op.calls", true, expectedOperationTag);
 
                 if (!disableLatencyMeter) {
+                    Tag expectedRequestTag = Tag.of(TagName.RequestStatusCode.toString(), "201/0");
                     this.validateMetrics(
                         expectedOperationTag,
-                        Tag.of(TagName.RequestStatusCode.toString(), "201/0"),
+                        expectedRequestTag,
                         1,
                         300
                     );
 
+                    // also ensure the replicaId dimension is populated for DIRECT mode
+                    if (this.client.asyncClient().getConnectionPolicy().getConnectionMode() == ConnectionMode.DIRECT) {
+                        Meter foundMeter = this.assertMetrics(
+                            "cosmos.client.req.rntbd.latency",
+                            true,
+                            expectedRequestTag);
+                        assertThat(foundMeter).isNotNull();
+                        boolean replicaIdDimensionExists = foundMeter
+                            .getId()
+                            .getTags()
+                            .stream()
+                            .anyMatch(tag -> tag.getKey().equals(TagName.ReplicaId.toString()) &&
+                                !tag.getValue().equals("NONE"));
+                        assertThat(replicaIdDimensionExists)
+                            .isEqualTo(true);
+                    }
+
                     this.validateMetrics(
                         Tag.of(
                             TagName.Operation.toString(), "Document/Create"),
@@ -1095,6 +1119,11 @@ public void meterTagNameFromStringConversion() {
             .isSameAs(CosmosMetricTagName.SERVICE_ADDRESS);
         assertThat(CosmosMetricTagName.fromString("serviceEndpoint"))
             .isSameAs(CosmosMetricTagName.SERVICE_ENDPOINT);
+        assertThat(CosmosMetricTagName.fromString("partitionID"))
+            .isSameAs(CosmosMetricTagName.PARTITION_ID);
+        assertThat(CosmosMetricTagName.fromString("REPLICAid"))
+            .isSameAs(CosmosMetricTagName.REPLICA_ID);
+
     }
 
     @Test(groups = {"simple"}, timeOut = TIMEOUT)

diff --git a/sdk/cosmos/azure-cosmos/CHANGELOG.md b/sdk/cosmos/azure-cosmos/CHANGELOG.md
@@ -4,7 +4,7 @@
 
 #### Features Added
 * Added the capability to filter request-level metrics based on diagnostic thresholds. Request-level metrics usually are used to capture metrics per backend endpoint/replica - a high cardinality dimension. Filtering by diagnostic thresholds reduces the overhead - but also means request-level metrics can only be used for debugging purposes - not for monitoring purposes. So, it is important to use the unfiltered operation-level metrics for health monitoring in this case. - See [PR 35114](https://github.com/Azure/azure-sdk-for-java/pull/35114)
-
+* Added optional tags/dimensions for PartitionId/ReplicaId as alternative to ServiceAddress for direct-mode (rntbd) request-level metrics. - See [PR 35164](https://github.com/Azure/azure-sdk-for-java/pull/35164) 
 #### Breaking Changes
 
 #### Bugs Fixed