2.22.1

Changes between 2.22.0 and 2.22.1

This is a patch release with a log-related bug fix. It is backward compatible with 2.22.0. All users are encouraged to use it.

Remove default debug setting

GitHub issue: #771

2.22.0

Changes between 2.21.0 and 2.22.0

This is a minor release with new features, usability improvements, and dependency upgrades. It is backward compatible with 2.21.0. All users are encouraged to use it.

Delay consumer recovery after cancel

GitHub PR: #719

Set default size of TCP buffers to 65,536

GitHub issue: #703

Add -tnd to control TCP_NODELAY

GitHub PR: #693

Add options to configure size of TCP buffers

GitHub issue: #689

Bump dependencies

GitHub issue: #690

2.21.0

Changes between 2.20.0 and 2.21.0

This is a minor release with a usability improvement and dependency upgrades. It is backward compatible with 2.20.0. All users are encouraged to use it.

Refuse to start with --multi-ack-every > --qos

GitHub issue: #614

Bump dependencies

GitHub issue: #576

2.20.0

Changes between 2.19.0 and 2.20.0

This is a minor release with new features, usability improvements, bug fixes, and dependency upgrades. It is backward compatible with 2.19.0. All users are encouraged to use it.

Use Java 21 in Docker image

GitHub PR: #573

Add confirm size metrics

GitHub PR: #531

Fix NIO thread pool size calculation

GitHub PR: #528

Document defaults

GitHub PR: #527

Verbose mode for functionality tests

GitHub issue: #515

Support OAuth2 authorization flow

GitHub PR: #476

Support multi-instance synchronized start-up

GitHub PR: #474

Do not crash on timestamp extraction

GitHub PR: #464

Use durable exchanges

GitHub PR: #460

Start metrics service before consumers and publishers

GitHub issue: #459

Bump dependencies

GitHub issue: #458

2.19.0

Changes between 2.18.0 and 2.19.0

This is a minor release with new features, usability improvements, bug fixes, and dependency upgrades. It is backward compatible with 2.18.0. All users are encouraged to use it.

This release introduces a new way to run PerfTest with a single JAR file (uber JAR). This is as of now the preferred way to run PerfTest. The archives (ZIP and TAR.GZ) and the native executable are deprecated and will be removed in PerfTest 3.0.

Deprecate native executable

GitHub issue: #449

PerfTestMulti: Add benchmark specification parameter for quorum queues

GitHub issue: #448

Install Ubuntu updates when building the image

GitHub PR: #426

JAVA_OPTS doesn't work the way that the docs indicate

GitHub issue: #416

Add compact console output format

GitHub PR: #415

Remove legacy metrics

GitHub PR: #414

Refactor metrics collection and output

GitHub PR: #413

Allow several values for --queue-args

GitHub PR: #411

Initialize globals metrics once producers/consumer are started

GitHub PR: #408

Use atomic utilities instead of synchronization in stats

GitHub PR: #407

Set 1s random start delay when producer rate is set

GitHub PR: #406

Use Guava rate limiter instead of internal solution

GitHub PR: #405

Use shutdown hook to print stats summary

GitHub PR: #402

Use millisecond granularity for publisher ramp-up period

GitHub PR: #401

Add latency metrics in final metrics reporting

GitHub issue: #400

Convert x-single-active-consumer to a bool

GitHub PR: #397

Bump dependencies

GitHub issue: #367

Use Java HttpServer to expose metrics and monitoring

GitHub issue: #358

Use uber JAR for packaging

GitHub issue: #351

Fix elapsed time in PerfTestMulti report

GitHub issue: #453

2.19.0.RC3

Changes between 2.18.0 and 2.19.0.RC3

This is a pre-release with new features, usability improvements, bug fixes, and dependency upgrades. It is backward compatible with 2.18.0. All users are encouraged to test it.

This release introduces a new way to run PerfTest with a single JAR file (uber JAR). This is as of now the preferred way to run PerfTest. The archives (ZIP and TAR.GZ) and the native executable are deprecated and will be removed in PerfTest 3.0.

Deprecate native executable

GitHub issue: #449

PerfTestMulti: Add benchmark specification parameter for quorum queues

GitHub issue: #448

Install Ubuntu updates when building the image

GitHub PR: #426

JAVA_OPTS doesn't work the way that the docs indicate

GitHub issue: #416

Add compact console output format

GitHub PR: #415

Remove legacy metrics

GitHub PR: #414

Refactor metrics collection and output

GitHub PR: #413

Allow several values for --queue-args

GitHub PR: #411

Initialize globals metrics once producers/consumer are started

GitHub PR: #408

Use atomic utilities instead of synchronization in stats

GitHub PR: #407

Set 1s random start delay when producer rate is set

GitHub PR: #406

Use Guava rate limiter instead of internal solution

GitHub PR: #405

Use shutdown hook to print stats summary

GitHub PR: #402

Use millisecond granularity for publisher ramp-up period

GitHub PR: #401

Add latency metrics in final metrics reporting

GitHub issue: #400

Convert x-single-active-consumer to a bool

GitHub PR: #397

Bump dependencies

GitHub issue: #367

Use Java HttpServer to expose metrics and monitoring

GitHub issue: #358

Use uber JAR for packaging

GitHub issue: #351

Fix elapsed time in PerfTestMulti report

GitHub issue: #453

2.19.0.RC2

Changes between 2.18.0 and 2.19.0.RC2

This is a pre-release with new features, usability improvements, bug fixes, and dependency upgrades. It is backward compatible with 2.18.0. All users are encouraged to test it.

This release introduces a new way to run PerfTest with a single JAR file (uber JAR). This is as of now the preferred way to run PerfTest. The archives (ZIP and TAR.GZ) and the native executable are deprecated and will be removed in PerfTest 3.0.

Deprecate native executable

GitHub issue: #449

PerfTestMulti: Add benchmark specification parameter for quorum queues

GitHub issue: #448

Install Ubuntu updates when building the image

GitHub PR: #426

JAVA_OPTS doesn't work the way that the docs indicate

GitHub issue: #416

Add compact console output format

GitHub PR: #415

Remove legacy metrics

GitHub PR: #414

Refactor metrics collection and output

GitHub PR: #413

Allow several values for --queue-args

GitHub PR: #411

Initialize globals metrics once producers/consumer are started

GitHub PR: #408

Use atomic utilities instead of synchronization in stats

GitHub PR: #407

Set 1s random start delay when producer rate is set

GitHub PR: #406

Use Guava rate limiter instead of internal solution

GitHub PR: #405

Use shutdown hook to print stats summary

GitHub PR: #402

Use millisecond granularity for publisher ramp-up period

GitHub PR: #401

Add latency metrics in final metrics reporting

GitHub issue: #400

Convert x-single-active-consumer to a bool

GitHub PR: #397

Bump dependencies

GitHub issue: #367

Use Java HttpServer to expose metrics and monitoring

GitHub issue: #358

Use uber JAR for packaging

GitHub issue: #351

2.19.0.RC1

Changes between 2.18.0 and 2.19.0.RC1

This is a pre-release with usability improvements and dependency upgrades. It is backward compatible with 2.18.0. All users are encouraged to test it.

This release introduces a new way to run PerfTest with a single JAR file (uber JAR). This is as of now the preferred way and the archives (ZIP and TAR.GZ) will be deprecated soon.

Use Java HttpServer to expose metrics and monitoring

GitHub issue: #358

Use uber JAR for packaging

GitHub issue: #351

Bump dependencies

GitHub issue: #367

2.18.0

Changes between 2.17.0 and 2.18.0

This is a minor release with bug fixes and dependency upgrades. It is backward compatible with 2.17.0. All users are encouraged to upgrade.

Interpret --producer-random-start-delay as seconds in scheduling

GitHub issue: #341

Multi-ack can block consumption when rate limitation is enabled

GitHub issue: #332

Use Ubuntu 22.04 in Docker image

GitHub issue: #329

Some classic queues don't have consumers after node restart

GitHub issue: #330

Bump dependencies

GitHub issue: #322

2.18.0.RC2

Changes between 2.17.0 and 2.18.0.RC1

This is a pre-release with bug fixes and dependency upgrades. It is backward compatible with 2.17.0. All users are encouraged to test it.

Interpret --producer-random-start-delay as seconds in scheduling

GitHub issue: #341

Multi-ack can block consumption when rate limitation is enabled

GitHub issue: #332

Use Ubuntu 22.04 in Docker image

GitHub issue: #329

Some classic queues don't have consumers after node restart

GitHub issue: #330

Bump dependencies

GitHub issue: #322