Consistently retrying 449 client-side across Gateway modes by FabianMeiswinkel · Pull Request #49332 · Azure/azure-sdk-for-java

FabianMeiswinkel · 2026-06-01T10:18:47Z

Description

Summary

Unifies 449 (Retry With) handling across Gateway V1 and Gateway V2/Thin Client so retries are orchestrated client-side.

Adds a shared RetryWithRetryPolicy for 449 retry behavior.
Adds GatewayRetryWithRetryPolicy to combine gateway metadata retry handling with client-side 449 retries.
Sends x-ms-noretry-449 for Gateway V1 requests so the service does not drive the retry loop.
Leaves Gateway V2/Thin Client without the Gateway V1-only opt-out header.
Applies remaining retry budget and client retry metadata on gateway 449 retry attempts.
Treats 449 as an expected retryable response to avoid noisy error-level logs.
Reuses the shared 449 policy from direct connectivity’s GoneAndRetryWithRetryPolicy.

Testing

Added unit coverage for gateway 449 retry policy behavior.
Added Gateway V1 header coverage for x-ms-noretry-449.
Updated Gateway V1 and Gateway V2 fault-injection coverage so injected 449 responses are expected to retry client-side and validate the retry attempt uses the client-side retry budget.
Updated direct connectivity retry policy tests to use a real RetryWithException.

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.

… for empty pages

…202605 Update JarStorageAccountName from oltpsparkcijarstore0326 to oltpsparkcijarstore0529 to point to the new storage account provisioned in the current ephemeral tenant rotation. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…bianMeiswinkel/azure-sdk-for-java into users/fabianm/ReadManyByPKFix

…to users/fabianm/449Retry

FabianMeiswinkel · 2026-06-01T10:25:01Z

/azp run java - cosmos - tests

azure-pipelines · 2026-06-01T10:25:27Z

Azure Pipelines successfully started running 1 pipeline(s).

Copilot

Pull request overview

This PR updates the Cosmos Gateway request path to handle HTTP 449 (RetryWith) consistently via client-side retry, and adds a gateway header to disable the legacy Gateway V1 server-side 449 retry loop where applicable.

Changes:

Add a reusable RetryWithRetryPolicy and a new GatewayRetryWithRetryPolicy that combines 449 RetryWith retry + existing gateway metadata/network retry behavior.
Ensure gateway requests send x-ms-noretry-449: true so 449 is retried client-side instead of via Gateway V1 server-side looping (with a ThinClient override).
Update/add unit and fault-injection tests to validate 449 retry behavior and diagnostics.

Show a summary per file

File	Description
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/ThinClientStoreModel.java	Overrides gateway 449 header behavior for ThinClient mode.
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/RxGatewayStoreModel.java	Adds `x-ms-noretry-449` header and switches gateway invoke path to new composite retry policy w/ timeout propagation.
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/RetryWithRetryPolicy.java	New standalone retry policy implementing backoff + timeout budgeting for 449.
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/HttpConstants.java	Introduces `NO_RETRY_449` header constant.
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/GatewayRetryWithRetryPolicy.java	New composite retry policy for gateway (449 RetryWith + existing metadata/network retries).
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/Exceptions.java	Treats 449 as “commonly expected” to reduce noisy logs.
sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/directconnectivity/GoneAndRetryWithRetryPolicy.java	Refactors to use the new shared `RetryWithRetryPolicy` implementation.
sdk/cosmos/azure-cosmos-tests/src/test/java/com/azure/cosmos/implementation/RxGatewayStoreModelTest.java	Adds unit tests asserting `x-ms-noretry-449` header behavior.
sdk/cosmos/azure-cosmos-tests/src/test/java/com/azure/cosmos/implementation/GatewayRetryWithRetryPolicyTest.java	New unit tests for the composite gateway retry policy.
sdk/cosmos/azure-cosmos-tests/src/test/java/com/azure/cosmos/implementation/directconnectivity/GoneAndRetryWithRetryPolicyTest.java	Adjusts RetryWith test setup to match refactor.
sdk/cosmos/azure-cosmos-tests/src/test/java/com/azure/cosmos/faultinjection/FaultInjectionServerErrorRuleOnGatewayV2Tests.java	Updates expectation: 449 is retried; validates retry diagnostics.
sdk/cosmos/azure-cosmos-tests/src/test/java/com/azure/cosmos/faultinjection/FaultInjectionServerErrorRuleOnGatewayTests.java	Updates expectation: 449 is retried; validates retry diagnostics.

Copilot's findings

Files reviewed: 13/13 changed files
Comments generated: 2

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

FabianMeiswinkel · 2026-06-01T10:35:49Z

/azp run java - cosmos - tests

azure-pipelines · 2026-06-01T10:36:10Z

Azure Pipelines successfully started running 1 pipeline(s).

FabianMeiswinkel · 2026-06-01T10:56:45Z

/azp run java - cosmos - tests

azure-pipelines · 2026-06-01T10:57:08Z

Azure Pipelines successfully started running 1 pipeline(s).

azure-pipelines · 2026-06-02T09:42:49Z

Azure Pipelines successfully started running 1 pipeline(s).

FabianMeiswinkel · 2026-06-02T11:48:55Z

/azp run java - cosmos - tests

FabianMeiswinkel · 2026-06-02T11:49:08Z

/azp run java - cosmos - spark

azure-pipelines · 2026-06-02T11:49:19Z

Azure Pipelines successfully started running 1 pipeline(s).

azure-pipelines · 2026-06-02T11:49:21Z

Azure Pipelines successfully started running 1 pipeline(s).

FabianMeiswinkel · 2026-06-02T12:13:56Z

/azp run java - cosmos - kafka

azure-pipelines · 2026-06-02T12:14:07Z

Azure Pipelines successfully started running 1 pipeline(s).

FabianMeiswinkel · 2026-06-02T19:40:57Z

/azp run java - cosmos - kafka

FabianMeiswinkel · 2026-06-02T19:41:07Z

/azp run java - cosmos - spark

azure-pipelines · 2026-06-02T19:41:15Z

Azure Pipelines successfully started running 1 pipeline(s).

FabianMeiswinkel · 2026-06-02T19:41:17Z

/azp run java - cosmos - tests

azure-pipelines · 2026-06-02T19:41:17Z

Azure Pipelines successfully started running 1 pipeline(s).

azure-pipelines · 2026-06-02T19:41:35Z

Azure Pipelines successfully started running 1 pipeline(s).

FabianMeiswinkel · 2026-06-02T21:59:56Z

/azp run java - cosmos - kafka

FabianMeiswinkel · 2026-06-02T22:00:04Z

/azp run java - cosmos - tests

azure-pipelines · 2026-06-02T22:00:14Z

Azure Pipelines successfully started running 1 pipeline(s).

FabianMeiswinkel · 2026-06-02T22:00:14Z

/azp run java - cosmos - spark

azure-pipelines · 2026-06-02T22:00:25Z

Azure Pipelines successfully started running 1 pipeline(s).

azure-pipelines · 2026-06-02T22:00:30Z

Azure Pipelines successfully started running 1 pipeline(s).

FabianMeiswinkel and others added 11 commits May 29, 2026 14:39

Fixes UnsupportedOperationException when using readManyByPartitionKey…

7173494

… for empty pages

Updated changelogs

a442db3

Merge branch 'main' into users/fabianm/ReadManyByPKFix

3ba04b0

Merge branch 'upstream-main' into users/fabianm/ReadManyByPKFix

af2b89d

Keeping FeedResponse.header final

671d37d

Merge branch 'users/fabianm/ReadManyByPKFix' of https://github.com/Fa…

ef4ac11

…bianMeiswinkel/azure-sdk-for-java into users/fabianm/ReadManyByPKFix

Merge branch 'main' into users/fabianm/ReadManyByPKFix

1766d57

Unifying 449 to be retried client-side across Gateway modes

03be254

Honoring remaining timeout

4169fbc

Merge branch 'main' of https://github.com/Azure/azure-sdk-for-java in…

352d523

…to users/fabianm/449Retry

Copilot AI review requested due to automatic review settings June 1, 2026 10:18

FabianMeiswinkel requested review from a team and kirankumarkolli as code owners June 1, 2026 10:18

Copilot started reviewing on behalf of FabianMeiswinkel June 1, 2026 10:19 View session

github-actions Bot added the Cosmos label Jun 1, 2026

Update CHANGELOG.md

4487103

Copilot AI reviewed Jun 1, 2026

View reviewed changes

Comment thread .../azure-cosmos/src/main/java/com/azure/cosmos/implementation/GatewayRetryWithRetryPolicy.java

Comment thread sdk/cosmos/azure-cosmos/src/main/java/com/azure/cosmos/implementation/RxGatewayStoreModel.java

Potential fix for pull request finding

3cdbaa7

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Update GatewayRetryWithRetryPolicy.java

2542491

jeet1995 reviewed Jun 1, 2026

View reviewed changes

Comment thread ...ain/java/com/azure/cosmos/implementation/directconnectivity/GoneAndRetryWithRetryPolicy.java

jeet1995 approved these changes Jun 1, 2026

View reviewed changes

Fixing test issue

ff30311

Fixing tests issues

fdfa730

FabianMeiswinkel added 2 commits June 2, 2026 19:39

Update QueryValidationTests.java

070a5bf

Merge branch 'main' into users/fabianm/449Retry

bc11211

jeet1995 approved these changes Jun 2, 2026

View reviewed changes

kushagraThapar approved these changes Jun 2, 2026

View reviewed changes

FabianMeiswinkel added 2 commits June 2, 2026 21:59

Fixing flaky query tests due to container creation being async

ccabf82

Merge branch 'main' into users/fabianm/449Retry

56fda90

jeet1995 approved these changes Jun 2, 2026

View reviewed changes

FabianMeiswinkel enabled auto-merge (squash) June 3, 2026 02:00

saragluna approved these changes Jun 3, 2026

View reviewed changes

FabianMeiswinkel merged commit 85fadfc into Azure:main Jun 3, 2026
139 checks passed

Conversation

FabianMeiswinkel commented Jun 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Summary

Testing

All SDK Contribution checklist:

General Guidelines and Best Practices

Testing Guidelines

Uh oh!

FabianMeiswinkel commented Jun 1, 2026

Uh oh!

azure-pipelines Bot commented Jun 1, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Copilot's findings

Uh oh!

Uh oh!

Uh oh!

FabianMeiswinkel commented Jun 1, 2026

Uh oh!

azure-pipelines Bot commented Jun 1, 2026

Uh oh!

FabianMeiswinkel commented Jun 1, 2026

Uh oh!

azure-pipelines Bot commented Jun 1, 2026

Uh oh!

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

FabianMeiswinkel commented Jun 2, 2026

Uh oh!

FabianMeiswinkel commented Jun 2, 2026

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

FabianMeiswinkel commented Jun 2, 2026

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

FabianMeiswinkel commented Jun 2, 2026

Uh oh!

FabianMeiswinkel commented Jun 2, 2026

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

FabianMeiswinkel commented Jun 2, 2026

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

FabianMeiswinkel commented Jun 2, 2026

Uh oh!

FabianMeiswinkel commented Jun 2, 2026

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

FabianMeiswinkel commented Jun 2, 2026

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

azure-pipelines Bot commented Jun 2, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

FabianMeiswinkel commented Jun 1, 2026 •

edited

Loading