Yield429 #47406

nehrao1 · 2025-12-01T17:02:54Z

Description

Please add an informative description that covers that changes made by the pull request and link all relevant issues.

If an SDK is being regenerated based on a new swagger spec, a link to the pull request containing these swagger spec changes has been included above.

All SDK Contribution checklist:

The pull request does not introduce [breaking changes]
CHANGELOG is updated for new features, bug fixes or other significant changes.
I have read the contribution guidelines.

General Guidelines and Best Practices

Title of the pull request is clear and informative.
There are a small number of commits, each of which have an informative message. This means that previously merged commits do not appear in the history of the PR. For more information on cleaning up the commits in your PR, see this page.

Testing Guidelines

Pull request includes test coverage for the included changes.

Copilot

Pull request overview

This pull request implements throttling (429) handling for quorum reads and write barriers in Azure Cosmos DB's strong consistency model. The changes enable the SDK to properly detect and respond when all replicas return 429 Too Many Requests responses during barrier operations.

Key changes:

Added QuorumThrottled result type and throttling detection logic in quorum read/write barrier flows
Introduced WaitForReadBarrierResult wrapper class to propagate throttled responses
Added new substatus code SERVER_WRITE_BARRIER_THROTTLED for timeout exceptions caused by throttling

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 11 comments.

Show a summary per file

File	Description
sdk/cosmos/azure-cosmos/src/main/java/module-info.java	Module descriptor file deleted (entire file removed)
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/directconnectivity/QuorumReader.java	Added `QuorumThrottled` enum value, introduced `WaitForReadBarrierResult` class, and updated read barrier logic to handle throttled responses
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/directconnectivity/ConsistencyWriter.java	Added throttling detection in write barrier with `AtomicBoolean` tracking and throws `RequestTimeoutException` when retries exhausted due to 429s
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/RequestTimeoutException.java	Added new constructor accepting `subStatusCode` parameter
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/HttpConstants.java	Added new substatus code constant `SERVER_WRITE_BARRIER_THROTTLED = 21013`
sdk/cosmos/azure-cosmos-tests/src/test/java/com/azure/cosmos/implementation/directconnectivity/ConsistencyWriterTest.java	Added commented-out test skeleton and unused imports
sdk/cosmos/azure-cosmos-tests/src/test/java/com/azure/cosmos/implementation/ConsistencyTestsStrong.java	Added new test file with C# syntax errors (not valid Java code)
sdk/cosmos/azure-cosmos-test/src/main/java/module-info.java	Module descriptor file deleted (entire file removed)

You can also share your feedback on Copilot code review for a chance to win a $100 gift card. Take the survey.

Copilot · 2025-12-01T17:14:02Z

...re-cosmos/src/main/java/com/azure/cosmos/implementation/directconnectivity/QuorumReader.java

                        return waitForObs.flatMap(
                            waitFor -> {
-                                if (!waitFor) {
+                                if (waitFor.isSuccess) {


Logic error: The condition should be if (!waitFor.isSuccess) to match the original behavior. The current code inverts the logic - it returns QuorumSelected when the barrier succeeds, but it should return QuorumSelected when the barrier fails (matching the original if (!waitFor) condition).

Suggested change

if (waitFor.isSuccess) {

if (!waitFor.isSuccess) {

Copilot · 2025-12-01T17:14:02Z

.../src/test/java/com/azure/cosmos/implementation/directconnectivity/ConsistencyWriterTest.java

+import com.azure.cosmos.rx.TestSuiteBase;
 import io.reactivex.subscribers.TestSubscriber;
 import org.mockito.MockedStatic;
 import org.mockito.Mockito;
 import org.mockito.invocation.InvocationOnMock;
 import org.testng.annotations.DataProvider;
+import org.testng.annotations.Factory;


The unused imports com.azure.cosmos.rx.TestSuiteBase and org.testng.annotations.Factory should be removed.

Suggested change

import com.azure.cosmos.rx.TestSuiteBase;

import io.reactivex.subscribers.TestSubscriber;

import org.mockito.MockedStatic;

import org.mockito.Mockito;

import org.mockito.invocation.InvocationOnMock;

import org.testng.annotations.DataProvider;

import org.testng.annotations.Factory;

import io.reactivex.subscribers.TestSubscriber;

import org.mockito.MockedStatic;

import org.mockito.Mockito;

import org.mockito.invocation.InvocationOnMock;

import org.testng.annotations.DataProvider;

Copilot · 2025-12-01T17:14:03Z

.../src/test/java/com/azure/cosmos/implementation/directconnectivity/ConsistencyWriterTest.java

+    /*@Test(groups = "unit")
+    public void ConsistencyWriter_BarrierRetriesExhaustedWith429_Throws408Async() {
+        int barrierRequestCount = 0;
+
+        sessionContainer = Mockito.mock(ISessionContainer.class);
+
+        transportClient = Mockito.mock(TransportClient.class);
+        when(transportClient.invokeStoreAsync(
+                Mockito.any(Uri.class),
+                Mockito.any(RxDocumentServiceRequest.class))
+            .thenAnswer(i -> i.getArguments()[0]));
+    }*/
+


Commented-out code should be removed rather than left in the codebase. If this test is meant to be implemented later, create a separate issue or TODO comment explaining the intent.

Suggested change

/*@Test(groups = "unit")

public void ConsistencyWriter_BarrierRetriesExhaustedWith429_Throws408Async() {

int barrierRequestCount = 0;

sessionContainer = Mockito.mock(ISessionContainer.class);

transportClient = Mockito.mock(TransportClient.class);

when(transportClient.invokeStoreAsync(

Mockito.any(Uri.class),

Mockito.any(RxDocumentServiceRequest.class))

.thenAnswer(i -> i.getArguments()[0]));

}*/

Copilot · 2025-12-01T17:14:03Z

...smos/src/main/java/com/azure/cosmos/implementation/directconnectivity/ConsistencyWriter.java

+                    //if (responses != null && responses.stream().anyMatch(response -> response.globalCommittedLSN >= selectedGlobalCommittedLsn)) {
+                    //    return Mono.just(Boolean.TRUE);
+                    //}
+


Commented-out code should be removed. This appears to be dead code that was replaced by the new logic below. Keeping commented code reduces readability and maintainability.

Suggested change

//if (responses != null && responses.stream().anyMatch(response -> response.globalCommittedLSN >= selectedGlobalCommittedLsn)) {

// return Mono.just(Boolean.TRUE);

//}

Copilot · 2025-12-01T17:14:03Z

...smos/src/main/java/com/azure/cosmos/implementation/directconnectivity/ConsistencyWriter.java

+
+                    if (responses != null) {
+                        // Check if all replicas returned 429, but don't exit early.
+                        if (responses.stream().count() > 0 && responses.stream().allMatch(response -> response.getStoreResponse().getStatus() == HttpConstants.StatusCodes.TOO_MANY_REQUESTS)) {


The condition responses.stream().count() > 0 is redundant. If you're calling allMatch() on an empty stream, it returns true (vacuous truth). Consider using !responses.isEmpty() for clarity and better performance, or rely on the allMatch() predicate alone if an empty response set should not be treated as throttled.

Suggested change

if (responses.stream().count() > 0 && responses.stream().allMatch(response -> response.getStoreResponse().getStatus() == HttpConstants.StatusCodes.TOO_MANY_REQUESTS)) {

if (!responses.isEmpty() && responses.stream().allMatch(response -> response.getStoreResponse().getStatus() == HttpConstants.StatusCodes.TOO_MANY_REQUESTS)) {

Copilot · 2025-12-01T17:14:04Z

...re-cosmos/src/main/java/com/azure/cosmos/implementation/directconnectivity/QuorumReader.java

+    private class WaitForReadBarrierResult {
+        private final boolean isSuccess;
+        private final StoreResponse throttledResponse;
+
+        private WaitForReadBarrierResult(boolean isSuccess, StoreResponse throttledResponse) {
+            this.isSuccess = isSuccess;
+            this.throttledResponse = throttledResponse;
+        }


Missing JavaDoc documentation for the new inner class WaitForReadBarrierResult and its fields. According to Azure SDK guidelines, all public and internal APIs should include comprehensive JavaDoc explaining the purpose and usage.

Copilot · 2025-12-01T17:14:04Z

...smos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/RequestTimeoutException.java

+    /**
+     * Instantiates a new Request timeout exception.
+     */
+    public RequestTimeoutException(int subStatusCode) {
+        this(RMResources.RequestTimeout, null, subStatusCode);
+    }


Missing JavaDoc for the new constructor. The constructor should document the subStatusCode parameter to explain its purpose and valid values.

Copilot · 2025-12-01T17:14:05Z

.../src/test/java/com/azure/cosmos/implementation/directconnectivity/ConsistencyWriterTest.java


 import com.azure.cosmos.BridgeInternal;
 import com.azure.cosmos.ConsistencyLevel;
+import com.azure.cosmos.CosmosClientBuilder;


The unused import com.azure.cosmos.CosmosClientBuilder should be removed.

Suggested change

import com.azure.cosmos.CosmosClientBuilder;

Copilot · 2025-12-01T17:14:05Z

.../src/test/java/com/azure/cosmos/implementation/directconnectivity/ConsistencyWriterTest.java

 import static com.azure.cosmos.implementation.TestUtils.mockDiagnosticsClientContext;
 import static com.azure.cosmos.implementation.TestUtils.mockDocumentServiceRequest;
 import static org.assertj.core.api.Assertions.assertThat;
+import static org.mockito.Mockito.when;


The when static import from Mockito is unused in the visible code (it appears only in the commented-out test). If the commented test is removed, this import should also be removed.

Suggested change

import static org.mockito.Mockito.when;

Copilot · 2025-12-01T17:14:05Z

...azure-cosmos-tests/src/test/java/com/azure/cosmos/implementation/ConsistencyTestsStrong.java

+    public void injectTooManyRequestsFaultAndVerify429Count(boolean shouldRegionScopedSessionContainerEnabled) throws InterruptedException {
+        string databaseName = "newdatabase";
+
+        // Create a fault injection rule for TooManyRequests (429) in direct mode
+        FaultInjectionRule tooManyRequestsRule = new FaultInjectionRuleBuilder(
+            id: "TooManyRequestsRule-" + Guid.NewGuid(),
+            condition: new FaultInjectionConditionBuilder()
+            .WithOperationType(FaultInjectionOperationType.ReadItem)
+            .Build(),
+            result: FaultInjectionResultBuilder.GetResultBuilder(FaultInjectionServerErrorType.TooManyRequests)
+            .Build())
+        .Build();
+
+        // inject 404/1002 into all regions
+        FaultInjectionRule tooManyRequestsRule = new FaultInjectionRuleBuilder("tooManyRequestsRule-" + UUID.randomUUID())
+            .condition(new FaultInjectionConditionBuilder())
+            .withOperationType(FaultInjectionOperationType.READ_ITEM)
+            .result(
+                FaultInjectionResultBuilders
+                    .getResultBuilder(FaultInjectionServerErrorType.READ_SESSION_NOT_AVAILABLE)
+                    .build())
+            .build();
+
+        // Initialize the fault injector
+        FaultInjector faultInjector = new FaultInjector(new List<FaultInjectionRule> { tooManyRequestsRule });
+
+        CosmosClient cosmosClient = new CosmosClientBuilder(
+            connectionString: "")
+                    .WithConnectionModeDirect().WithFaultInjection(faultInjector)
+            .WithConsistencyLevel(Cosmos.ConsistencyLevel.Eventual)
+            .WithThrottlingRetryOptions(
+                maxRetryWaitTimeOnThrottledRequests: TimeSpan.FromSeconds(5), // Maximum wait time for retries
+            maxRetryAttemptsOnThrottledRequests: 2) // Maximum retry attempts
+                    .Build();
+
+
+        ContainerProperties containerProperties = new ContainerProperties(
+            id: "test",
+            partitionKeyPath: "/id");
+
+
+        // Create database and container
+        await cosmosClient.CreateDatabaseIfNotExistsAsync(databaseName);
+        Container container = await cosmosClient.GetDatabase(databaseName).CreateContainerIfNotExistsAsync(containerProperties);
+
+        dynamic testObject = new
+        {
+            id = Guid.NewGuid().ToString(),
+                Company = "Microsoft",
+                State = "WA"
+
+        };
+
+
+        await container.CreateItemAsync<dynamic>(testObject);
+        try
+        {
+            // Attempt to read the item
+            ItemResponse<dynamic> itemResponse = await container.ReadItemAsync<dynamic>(
+            testObject.id,
+            new Cosmos.PartitionKey(testObject.id));
+
+            // Print diagnostics
+            Console.WriteLine("Diagnostics:");
+            Console.WriteLine(itemResponse.Diagnostics.ToString());
+        }
+
+        catch (CosmosException ex)
+        {
+            // Handle other Cosmos exceptions
+            Console.WriteLine($"CosmosException: {ex.StatusCode} - {ex.Message}");
+            Console.WriteLine("Diagnostics:");
+            Console.WriteLine(ex.Diagnostics.ToString());
+        }
+        long hitCount = tooManyRequestsRule.GetHitCount();
+        Console.WriteLine($"Total 429 responses: {hitCount}");


This test file contains syntax errors and uses C# syntax instead of Java:

Line 32: string databaseName should be String databaseName

Lines 36-42, 45-52: Invalid syntax mixing C# and Java

Line 55: new List<FaultInjectionRule> should be Java syntax

Lines 57-64: Using C# named parameters syntax (connectionString:, WithConnectionModeDirect())

Lines 73-74, 85-91, 94-95: await keyword is C# syntax (Java uses .block() for reactive code)

Line 78: Guid.NewGuid() is C#, should be UUID.randomUUID()

Lines 100-103: Console.WriteLine is C#, should be System.out.println or logger

This entire test file needs to be rewritten using proper Java syntax and Reactor patterns.

Suggested change

public void injectTooManyRequestsFaultAndVerify429Count(boolean shouldRegionScopedSessionContainerEnabled) throws InterruptedException {

string databaseName = "newdatabase";

// Create a fault injection rule for TooManyRequests (429) in direct mode

FaultInjectionRule tooManyRequestsRule = new FaultInjectionRuleBuilder(

id: "TooManyRequestsRule-" + Guid.NewGuid(),

condition: new FaultInjectionConditionBuilder()

.WithOperationType(FaultInjectionOperationType.ReadItem)

.Build(),

result: FaultInjectionResultBuilder.GetResultBuilder(FaultInjectionServerErrorType.TooManyRequests)

.Build())

.Build();

// inject 404/1002 into all regions

FaultInjectionRule tooManyRequestsRule = new FaultInjectionRuleBuilder("tooManyRequestsRule-" + UUID.randomUUID())

.condition(new FaultInjectionConditionBuilder())

.withOperationType(FaultInjectionOperationType.READ_ITEM)

.result(

FaultInjectionResultBuilders

.getResultBuilder(FaultInjectionServerErrorType.READ_SESSION_NOT_AVAILABLE)

.build())

.build();

// Initialize the fault injector

FaultInjector faultInjector = new FaultInjector(new List<FaultInjectionRule> { tooManyRequestsRule });

CosmosClient cosmosClient = new CosmosClientBuilder(

connectionString: "")

.WithConnectionModeDirect().WithFaultInjection(faultInjector)

.WithConsistencyLevel(Cosmos.ConsistencyLevel.Eventual)

.WithThrottlingRetryOptions(

maxRetryWaitTimeOnThrottledRequests: TimeSpan.FromSeconds(5), // Maximum wait time for retries

maxRetryAttemptsOnThrottledRequests: 2) // Maximum retry attempts

.Build();

ContainerProperties containerProperties = new ContainerProperties(

id: "test",

partitionKeyPath: "/id");

// Create database and container

await cosmosClient.CreateDatabaseIfNotExistsAsync(databaseName);

Container container = await cosmosClient.GetDatabase(databaseName).CreateContainerIfNotExistsAsync(containerProperties);

dynamic testObject = new

{

id = Guid.NewGuid().ToString(),

Company = "Microsoft",

State = "WA"

};

await container.CreateItemAsync<dynamic>(testObject);

try

{

// Attempt to read the item

ItemResponse<dynamic> itemResponse = await container.ReadItemAsync<dynamic>(

testObject.id,

new Cosmos.PartitionKey(testObject.id));

// Print diagnostics

Console.WriteLine("Diagnostics:");

Console.WriteLine(itemResponse.Diagnostics.ToString());

}

catch (CosmosException ex)

{

// Handle other Cosmos exceptions

Console.WriteLine($"CosmosException: {ex.StatusCode} - {ex.Message}");

Console.WriteLine("Diagnostics:");

Console.WriteLine(ex.Diagnostics.ToString());

}

long hitCount = tooManyRequestsRule.GetHitCount();

Console.WriteLine($"Total 429 responses: {hitCount}");

public void injectTooManyRequestsFaultAndVerify429Count() throws InterruptedException {

String databaseName = "newdatabase";

// Create a fault injection rule for TooManyRequests (429) in direct mode

FaultInjectionRule tooManyRequestsRule = new FaultInjectionRuleBuilder("TooManyRequestsRule-" + UUID.randomUUID())

.condition(new FaultInjectionConditionBuilder()

.operationType(FaultInjectionOperationType.READ_ITEM)

.build())

.result(

FaultInjectionResultBuilders

.getResultBuilder(FaultInjectionServerErrorType.TOO_MANY_REQUESTS)

.build())

.build();

// Initialize the fault injector

List<FaultInjectionRule> rules = new ArrayList<>();

rules.add(tooManyRequestsRule);

FaultInjector faultInjector = new FaultInjector(rules);

CosmosClient cosmosClient = new CosmosClientBuilder()

.connectionString("") // TODO: Provide valid connection string

.directMode()

.faultInjection(faultInjector)

.consistencyLevel(ConsistencyLevel.EVENTUAL)

.throttlingRetryOptions(new ThrottlingRetryOptions()

.setMaxRetryWaitTime(5)

.setMaxRetryAttemptsOnThrottledRequests(2))

.buildClient();

ContainerProperties containerProperties = new ContainerProperties("test", "/id");

// Create database and container

cosmosClient.createDatabaseIfNotExists(databaseName).block();

CosmosContainer container = cosmosClient.getDatabase(databaseName)

.createContainerIfNotExists(containerProperties).block();

// Create a test object as a POJO or Map

String itemId = UUID.randomUUID().toString();

TestItem testObject = new TestItem(itemId, "Microsoft", "WA");

container.createItem(testObject).block();

try {

// Attempt to read the item

CosmosItemResponse<TestItem> itemResponse = container.readItem(

testObject.getId(),

new PartitionKey(testObject.getId()),

TestItem.class

).block();

// Print diagnostics

System.out.println("Diagnostics:");

System.out.println(itemResponse.getDiagnostics().toString());

} catch (CosmosException ex) {

// Handle other Cosmos exceptions

System.out.println("CosmosException: " + ex.getStatusCode() + " - " + ex.getMessage());

System.out.println("Diagnostics:");

System.out.println(ex.getDiagnostics().toString());

}

long hitCount = tooManyRequestsRule.getHitCount();

System.out.println("Total 429 responses: " + hitCount);

}

// Helper class for test item

static class TestItem {

private String id;

private String company;

private String state;

public TestItem(String id, String company, String state) {

this.id = id;

this.company = company;

this.state = state;

}

public String getId() {

return id;

}

public String getCompany() {

return company;

}

public String getState() {

return state;

}

nehrao1 added 2 commits November 25, 2025 10:50

consistency writer and quorum reader changes

e1f3891

adding tests

a993838

Copilot AI review requested due to automatic review settings December 1, 2025 17:02

nehrao1 requested review from a team and kirankumarkolli as code owners December 1, 2025 17:02

github-actions bot added the Cosmos label Dec 1, 2025

nehrao1 marked this pull request as draft December 1, 2025 17:03

Copilot started reviewing on behalf of nehrao1 December 1, 2025 17:03 View session

Copilot finished reviewing on behalf of nehrao1 December 1, 2025 17:05

Copilot AI reviewed Dec 1, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Yield429 #47406

Yield429 #47406

Uh oh!

nehrao1 commented Dec 1, 2025

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Copilot AI Dec 1, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

	//if (responses != null && responses.stream().anyMatch(response -> response.globalCommittedLSN >= selectedGlobalCommittedLsn)) {
	// return Mono.just(Boolean.TRUE);
	//}

	if (responses.stream().count() > 0 && responses.stream().allMatch(response -> response.getStoreResponse().getStatus() == HttpConstants.StatusCodes.TOO_MANY_REQUESTS)) {
	if (!responses.isEmpty() && responses.stream().allMatch(response -> response.getStoreResponse().getStatus() == HttpConstants.StatusCodes.TOO_MANY_REQUESTS)) {

Yield429 #47406

Are you sure you want to change the base?

Yield429 #47406

Uh oh!

Conversation

nehrao1 commented Dec 1, 2025

Description

All SDK Contribution checklist:

General Guidelines and Best Practices

Testing Guidelines

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Dec 1, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant