Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Initial draft

Table of Contents

This page is meant as a template for writing a KIP. To create a KIP choose Tools->Copy on this page and modify with your content and replace the heading with the next KIP number and a description of your issue. Replace anything in italics with your own description.

Status

Current state[One of "Under Discussion", "Accepted", "Rejected"]Draft

Discussion thread: here [Change the link from the KIP proposal email archive to your own email thread]

...

Please keep the discussion on the mailing list rather than commenting on the wiki (wiki discussions get unwieldy fast).

Motivation

Describe the problems you are trying to solve.

Public Interfaces

Briefly list any new interfaces that will be introduced as part of this proposal or any existing interfaces that will be removed or changed. The purpose of this section is to concisely call out the public contract that will come along with this feature.

A public interface is any change to the following:

  • Binary log format

  • The network protocol and api behavior

  • Any class in the public packages under clientsConfiguration, especially client configuration

    • org/apache/kafka/common/serialization

    • org/apache/kafka/common

    • org/apache/kafka/common/errors

    • org/apache/kafka/clients/producer

    • org/apache/kafka/clients/consumer (eventually, once stable)

  • Monitoring

  • Command line tools and arguments

  • Anything else that will likely break existing users in some way when they upgrade

Proposed Changes

Describe the new thing you want to do in appropriate detail. This may be fairly extensive and have large subsections of its own. Or it may be a few sentences. Use judgement based on the scope of the change.

Compatibility, Deprecation, and Migration Plan

  • What impact (if any) will there be on existing users?
  • If we are changing behavior how will we phase out the older behavior?
  • If we need special migration tools, describe them here.
  • When will we remove the existing behavior?

Test Plan

Describe in few sentences how the KIP will be tested. We are mostly interested in system tests (since unit-tests are specific to implementation details). How will we know that the implementation works as expected? How will we know nothing broke?

Rejected Alternatives

KIP-932: Queues for Kafka introduced share groups to Apache Kafka. This KIP introduces additional group configurations to give users more control over share groups.

Proposed Changes

Some of the configuration for share groups in KIP-932 is specified at the broker level without an equivalent group configuration.

First, it enables group configuration for the maximum number of delivery attempts, so error tolerance can be configured for each group. Second, it enables group configuration for the record lock limit per share-partition, so the balance between resource consumption and concurrency can be configured for each group. The defaults for the equivalent broker configurations were chosen to be suitable for most situations, but group overrides give administrators more control if they desire.

Finally, KIP-1222: Acquisition lock timeout renewal in share consumer explicit mode introduced renewal of acquisition locks. This enables processing of records to be extended indefinitely by repeatedly renewing the acquisition locks. A new group configuration is added to control whether this behavior is permitted for the group. If the value of the configuration is set to false, a renew acknowledgement will be rejected with error code INVALID_RECORD_STATE  and an error message which explains that renewing acquisition locks is not enabled for the group.

Public Interfaces

Configuration

Broker configuration

New configurations are added to specify the upper and lower bounds of the associated group configurations.

ConfigurationDescriptionValues

group.share.delivery.count.limit 

(Existing configuration)Default 5, minimum 2, maximum 10 (no change)

group.share.max.delivery.count.limit 

The maximum value of a group configuration for the maximum number of delivery attempts for a record delivered to a share group.

Default 10, minimum 5, maximum 25

group.share.min.delivery.count.limit 

The minimum value of a group configuration for the maximum number of delivery attempts for a record delivered to a share group.Default 2, minimum 2, maximum 5

group.share.partition.max.record.locks 

(Existing configuration)Default 2000, minimum 100, maximum 10000 (no change)

group.share.max.partition.max.record.locks 

The maximum value of a group configuration for the share-group record lock limit per share-partition.Default 4000, minimum 2000, maximum 10000
group.share.min.partition.max.record.locks The minimum value of a group configuration for the share-group record lock limit per share-partition.Default 100, minimum 100, maximum 2000

Group configuration

ConfigurationDescriptionValues
share.delivery.count.limit The maximum number of delivery attempts for a record delivered to the share group.

If not specified, uses the broker configuration group.share.delivery.count.limit .

If specified, minimum limited by the broker configuration group.share.min.delivery.count.limit , maximum limited by the broker configuration group.share.max.delivery.count.limit .

share.partition.max.record.locks Record lock limit per share-partition.

If not specified, uses the broker configuration group.share.partition.max.record.locks .

If specified, minimum limited by the broker configuration group.share.min.partitions.max.record.locks , maximum limited by the broker configuration group.share.max.partitions.max.record.locks .

share.renew.acknowledge.enable Whether the renew acknowledge type is enabled for the share group.

Valid values "true" (default) and "false"

Compatibility, Deprecation, and Migration Plan

The changes in this KIP are limited to adding a few new configurations to control existing behavior. There should be no compatibility concerns.

Test Plan

The feature will be thoroughly tested with unit and integration tests.

Rejected Alternatives

None so farIf there are alternative ways of accomplishing the same thing, what were they? The purpose of this section is to motivate why the design is the way it is and not some other way.