Status

Current state: Under discussion

Discussion thread: here

JIRA: https://issues.apache.org/jira/browse/KAFKA-16891

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

Motivation

The behavior of groups in Apache Kafka is more complicated and subtle than it first appears. To most users of Kafka, groups are synonymous with consumer groups. However, the "classic" consumer group protocol was extensible and there are several well-known extensions in use. For example, distributed workers in Kafka Connect also use groups as a coordination mechanism, and some applications such as schema registries have also built upon the consumer group protocol in interesting ways. These are all groups.

KIP-848 introduced the new consumer group protocol and modern consumer groups use this new protocol. KIP-932 introduces share groups, KIP-1071 introduces streams groups, and there may well be additional types of group in the future.

All of these types of groups share a namespace for group IDs, but the manner in which you administer a group and the operations you can perform upon it depend upon its type.

Here’s an example of the complexity. If you start up a distributed Kafka Connect worker using the default configuration, it creates a group called "connect-cluster" . This is a group, but it’s not a consumer group. You can’t see this group in the list of consumer groups with the kafka-consumer-groups.sh  tool, but if you try to describe a consumer group called "connect-cluster"  or even use this group ID with a consumer, you get an error.

Here's another example. If you use kafka-consumer-groups.sh --describe --group MYSHARE  where the group is a share group, the output is Error: Consumer group 'MYSHARE' does not exist . That's probably OK, but it's interesting to see how it gets there.

First, the admin client uses the ConsumerGroupDescribe RPC which responds with error code GROUP_ID_NOT_FOUND (69)  giving no indication that it found a group of the wrong type. Next, the admin client falls back to the pre-KIP-848 DescribeGroups RPC in case it's a classic consumer group.  This RPC succeeds and responds with error code NONE (0)  returning the group description with a status of Dead . It looks like a dead consumer group. Finally, this dead group is translated into the error message Error: Consumer group 'MYSHARE' does not exist . The output seems kind of acceptable, but the tool actually thinks it's dealing with a dead consumer group. It would be better if the ConsumerGroupDescribe RPC failed in a straightforward way.

This KIP tries to resolve some of these situations and make it easier to work out what’s going on with the groups on a cluster.

Proposed Changes

This KIP introduces a command-line tool called kafka-groups.sh  for displaying all of the groups and their types.

In situations where command-line tools are used to administer a group of the wrong type, the error message now indicates when the group type is wrong, rather than saying the group does not exist. This is achieved using a new error code in the Kafka protocol and a slight change in behavior of the admin client.

Listing groups

The KIP introduces a new tool called kafka-groups.sh to show all of the groups in a cluster, their types and the protocols they use. This lets you see consumer groups, share groups, Kafka Connect cluster groups, and any other custom groups all together. It doesn't replace the specific tools for the different types of group, but it does shed light on what's actually going on for administrators. Note that this does not require any changes to the Kafka protocol. The information is already available, but not directly accessible by the administrator.

The ListGroups RPC response returns three pieces of information for each group: group ID, type and protocol. For the common types of group, here is what they mean:

The new kafka-groups.sh tool makes all of this information available.

Describing groups using the admin client

A new exception InconsistentGroupTypeException  is used by Admin.describeConsumerGroups(Collection<String>) and Admin.describeShareGroups(Collection<String>)  to indicate that the group being described has the wrong type. This small change enables the error messages from the command line kafka-consumer-groups.sh  and kafka-share-groups.sh  to indicate when the group ID is found, but it has the wrong type.

Public Interfaces

Client API changes

Admin

Add the following methods on the org.apache.kafka.client.admin.Admin  interface.

Here are the method signatures:

   /**
    * Describe some classic groups in the cluster.
    *
    * @param groupIds The IDs of the groups to describe.
    * @param options  The options to use when describing the groups.
    * @return The DescribeClassicGroupsResult.
    */
   DescribeClassicGroupsResult describeClassicGroups(Collection<String> groupIds,
                                                     DescribeClassicGroupsOptions options);

   /**
    * Describe some classic groups in the cluster, with the default options.
    * <p>
    * This is a convenience method for {@link #describeClassicGroups(Collection, DescribeClassicGroupsOptions)}
    * with default options. See the overload for more details.
    *
    * @param groupIds The IDs of the groups to describe.
    * @return The DescribeClassicGroupsResult.
    */
   default DescribeClassicGroupsResult describeClassicGroups(Collection<String> groupIds) {
       return describeClassicGroups(groupIds, new DescribeClassicGroupsOptions());
   }

   /**
    * List the groups available in the cluster with the default options.
    *
    * <p>This is a convenience method for {@link #listGroups(ListGroupsOptions)} with default options.
    * See the overload for more details.
    *
    * @return The ListGroupsResult.
    */
   default ListGroupsResult listGroups() {
       return listGroups(new ListGroupsOptions());
   }
 
   /**
    * List the groups available in the cluster.
    *
    * @param options The options to use when listing the groups.
    * @return The ListGroupsResult.
    */
   ListGroupsResult listGroups(ListGroupsOptions options);

DescribeClassicGroupsOptions

/**
 * Options for {@link Admin#describeClassicGroups(Collection, DescribeClassicGroupsOptions)}.
 * <p>
 * The API of this class is evolving, see {@link Admin} for details.
 */
@InterfaceStability.Evolving
public class DescribeClassicGroupsOptions extends AbstractOptions<DescribeClassicGroupsOptions> {
    private boolean includeAuthorizedOperations;

    public DescribeClassicGroupsOptions includeAuthorizedOperations(boolean includeAuthorizedOperations) {
        this.includeAuthorizedOperations = includeAuthorizedOperations;
        return this;
    }

    public boolean includeAuthorizedOperations() {
        return includeAuthorizedOperations;
    }
}

DescribeClassicGroupsResult

package org.apache.kafka.client.admin;

/**
 * The result of the {@link KafkaAdminClient#describeClassicGroups(Collection, DescribeClassicGroupsOptions)}} call.
 *
 * The API of this class is evolving, see {@link Admin} for details.
 */
@InterfaceStability.Evolving
public class DescribeClassicGroupsResult {

    public DescribeClassicGroupsResult(final Map<String, KafkaFuture<ClassicGroupDescription>> futures);

    /**
     * Return a map from group id to futures which yield group descriptions.
     */
    public Map<String, KafkaFuture<ClassicGroupDescription>> describedGroups();

    /**
     * Return a future which yields all ClassicGroupDescription objects, if all the describes succeed.
     */
    public KafkaFuture<Map<String, ClassicGroupDescription>> all();
}

ClassicGroupDescription

package org.apache.kafka.client.admin;

/**
 * A detailed description of a single classic group in the cluster.
 */
@InterfaceStability.Evolving
public class ClassicGroupDescription {
    public ClassicGroupDescription(String groupId,
                                   String protocol,
                                   Collection<MemberDescription> members,
                                   String partitionAssignor,
                                   ClassicGroupState state,
                                   Node coordinator);

    public ClassicGroupDescription(String groupId,
                                   String protocol,
                                   Collection<MemberDescription> members,
                                   String partitionAssignor,
                                   ClassicGroupState state,
                                   Node coordinator,
                                   Set<AclOperation> authorizedOperations);

    /**
     * The id of the classic group.
     */
    public String groupId();

    /**
     * The group protocol type.
     */
    public String protocol();

    /**
     * If the group is a simple consumer group or not.
     */
    public boolean isSimpleConsumerGroup();

    /**
     * A list of the members of the classic group.
     */
    public Collection<MemberDescription> members();

    /**
     * The group partition assignor.
     */
    public String partitionAssignor();

    /**
     * The classic group state, or UNKNOWN if the state is too new for us to parse.
     */
    public ClassicGroupState state();

    /**
     * The group coordinator, or null if the coordinator is not known.
     */
    public Node coordinator();

    /**
     * authorizedOperations for this group, or null if that information is not known.
     */
    public Set<AclOperation> authorizedOperations();
}

ClassicGroupState

package org.apache.kafka.common;

/**
 * The classic group state.
 */
public enum ClassicGroupState {
    UNKNOWN("Unknown"),
    PREPARING_REBALANCE("PreparingRebalance"),
    COMPLETING_REBALANCE("CompletingRebalance"),
    STABLE("Stable"),
    DEAD("Dead"),
    EMPTY("Empty");

    ClassicGroupState(String name);

    /**
     * Case-insensitive classic group state lookup by string name.
     */
    public static ClassicGroupState parse(String name);

    public String toString();
}

ListGroupsOptions

package org.apache.kafka.client.admin;
 
import org.apache.kafka.common.GroupType;
 
/**
 * Options for {@link Admin#listGroups(ListGroupsOptions)}.
 *
 * The API of this class is evolving, see {@link Admin} for details.
 */
@InterfaceStability.Evolving
public class ListGroupsOptions extends AbstractOptions<ListGroupsOptions> {

    /**
     * If types is set, only groups of these types will be returned by listGroups().
     * Otherwise, all groups are returned.
     */
    public ListGroupsOptions withTypes(Set<GroupType> types) {
        this.types = (types == null || types.isEmpty()) ? Collections.emptySet() : new HashSet<>(types);
        return this;
    }

    /**
     * Returns the list of group types that are requested or empty if no types have been specified.
     */
    public Set<GroupType> types() {
        return types;
    }
}

ListGroupsResult

package org.apache.kafka.clients.admin;
   
/**
 * The result of the {@link Admin#listGroups(ListGroupsOptions)} call.
 * <p>
 * The API of this class is evolving, see {@link Admin} for details.
 */
@InterfaceStability.Evolving
public class ListGroupsResult {
    ListGroupsResult(KafkaFuture<Collection<Object>> future) {
        super(future);
    }

    /**
     * Returns a future that yields either an exception, or the full set of group listings.
     */
    public KafkaFuture<Collection<GroupListing>> all() {
    }
  
    /**
     * Returns a future which yields just the valid listings.
     */
    public KafkaFuture<Collection<GroupListing>> valid() {
    }
   
    /**
     * Returns a future which yields just the errors which occurred.
     */
    public KafkaFuture<Collection<Throwable>> errors() {
    }
}

GroupListing

package org.apache.kafka.client.admin;
  
/**
 * A listing of a group in the cluster.
 * <p>
 * The API of this class is evolving, see {@link Admin} for details.
 */
@InterfaceStability.Evolving
public class GroupListing {
  public GroupListing(String groupId, Optional<GroupType> type, String protocol);

  /**
   * The id of the group.
   */
  public String groupId();
  
  /**
   * The group type.
   */
  public Optional<GroupType> type();
  
  /**
   * The group protocol type.
   */
  public String protocol();

  /**
   * If the group is a simple consumer group or not.
   */
  public boolean isSimpleConsumerGroup();
}

ConsumerGroupListing

This class is modified to extend org.apache.kafka.clients.admin.GroupListing .

ShareGroupListing

This class is modified to extend org.apache.kafka.clients.admin.GroupListing .

Exceptions

The following new exception is added to the org.apache.kafka.common.errors  package corresponding to the new error code in the Kafka protocol.

• InconsistentGroupTypeException  - Indicates that the group exists but the group type is inconsistent with the operation.

Kafka protocol changes

This KIP adds the following error code to the Kafka protocol.

• INCONSISTENT_GROUP_TYPE  (value TBD) - Indicates that the group exists but the group type is inconsistent with the operation.

This error code is used when the following RPCs are used with an existing group of the wrong type:

• ConsumerGroupDescribe
• ShareGroupDescribe

These RPCs are used by administrative tools and the new error code will help with the usability of the tools. A new version of ConsumerGroupDescribe with no schema change will be required to support the new error code. Because ShareGroupDescribe is still unstable, this error code can be added to the v0 RPC.

The remaining RPCs which work with groups, such as OffsetFetch and TxnOffsetCommit, continue to fail with GROUP_ID_NOT_FOUND  if used against a group of the wrong type.

Command-line tools

kafka-groups.sh

A new tool called kafka-groups.sh  is introduced for listing groups of any kind. It has the following options:

Note that --consumer  actually matches all groups whose type is Consumer , and groups whose type is Classic and protocol type is "consumer", and also "simple" consumer groups whose type is Classic and protocol type is "" . The filtering is done in the kafka-groups.sh  tool.

Here are some examples.

To list all of the groups:

$ bin/kafka-groups.sh --bootstrap-server localhost:9092 --list
GROUP                   TYPE          PROTOCOL
old-consumer-group      Classic       consumer
new-consumer-group      Consumer      consumer
connect-cluster         Classic       connect
share-group             Share         share
schema-registry         Classic       sr
simple-consumer-group   Classic

To list all of the consumer groups, silently merging together all of the different kinds of consumer group:

$ bin/kafka-groups.sh --bootstrap-server localhost:9092 --list --consumer
GROUP                   TYPE          PROTOCOL
old-consumer-group      Classic       consumer
new-consumer-group      Consumer      consumer
simple-consumer-group   Classic

To list all of the KIP-848 consumer groups:

$ bin/kafka-groups.sh --bootstrap-server localhost:9092 --list --group-type consumer
GROUP                   TYPE          PROTOCOL
new-consumer-group      Consumer      consumer

To list all of the share groups:

$ bin/kafka-groups.sh --bootstrap-server localhost:9092 --list --share
GROUP                   TYPE          PROTOCOL
share-group             Share         share

kafka-consumer-groups.sh

For all operations which act on a single group, if that group exists but is not a consumer group, the command fails with a message indicating that the group type is incorrect, rather than the existing message that the group does not exist.

For example, if you try to describe a share group, the output will look like this:

$ bin/kafka-consumer-groups.sh --bootstrap-server localhost:9092 --describe --group SG1
Error: Group 'SG1' is not a consumer group.

kafka-share-groups.sh

For all operations which act on a single group, if that group exists but is not a share group, the command fails with a message indicating that the group type is incorrect, rather than the existing message that the group does not exist.

For example, if you try to describe a consumer group, the output will look like this:

$ bin/kafka-share-groups.sh --bootstrap-server localhost:9092 --describe --group CG1
Error: Group 'CG1' is not a share group.

Compatibility, Deprecation, and Migration Plan

ListGroups RPC

When writing this KIP, it seemed that perhaps the ListGroups RPC would need to be enhanced to create the Admin.listGroups()  method. This turned out not to be necessary.

ListGroups v5 introduces the TypesFilter  in the request and the GroupType  in the response. This KIP does not need to introduce a new version of the ListGroups RPC, because:

• If the broker does not support ListGroups v5, setting a types filter using ListGroupOptions.withTypes() results in UnsupportedVersionException when the ListGroups request is serialized. (This is KIP-848 behavior.)
• If the broker receives a group type in TypesFilter  that it does not support in a ListGroups v5 (or later) request, the filter is treated as unknown group type which thus matches no groups. (This is KIP-848 behavior.)
• If the client receives a group type that it does not understand in a ListGroups v5 (or later) response, the group type is GroupType.UNKNOWN . This is because the group type in the ListGroups RPC response cannot be parsed into a value of the GroupType  enumeration, and the default value of UNKNOWN  is used in these situations.

The existing behavior suffices and there is no compatibility problem.

DescribeConsumerGroups

Prior to this KIP, the behavior of Admin.describeConsumerGroups(Collection<String>) seems a little unusual when used with groups which are not consumer groups. You can describe a collection of group IDs, some of which might exist and others might not. Here's how the response is built:

1. If the group is a consumer group and the client is authorized to describe the group and there was no error, the group information is returned, along with the authorized operations if requested.
2. If the group is not a consumer group (either does not exist or wrong type) and the client is authorized to describe the group and there was no error, the group information for a dead group is returned, along with the authorized operations if requested.
3. If the client is not authorized to the describe the group, the group information error code is set to GROUP_AUTHORIZATION_FAILED . 
4. If there was an error describing the group, the group information error code is set.

In cases (1) and (2), the admin client considers the operation a success, and this means the KafkaFuture for this group completes successfully. In cases (3) and (4), the admin client considers the operation unsuccessful, and this means the KafkaFuture for this group completes exceptionally. Case (2) is the tricky one because you can't readily tell what the dead group means. This is why the admin tools use output like Error: Consumer group 'MYSHARE" does not exist, even when the group ID is recognised and it's just the wrong type.

After this KIP, if you use Admin.describeConsumerGroup(Collection<String>)  to describe a group which is not a consumer group, case (2) above will result in the group information error code set to INCONSISTENT_GROUP_TYPE . In the admin client, this means the future for this group completes exceptionally with InconsistentGroupTypeException  rather than succeeding with a dead consumer group.

Test Plan

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

Rejected Alternatives

It would be possible to preserve the current behavior of Admin.describeConsumerGroups(Collection<String>)  when used with a group which is not a consumer group and introduce an option to ask it to validate the group type rather than converting any indescribable group into a dead consumer group. This seems like an unnecessary complication with little benefit.

The error code INCONSISTENT_GROUP_PROTOCOL  already exists and is used with the classic consumer group protocol if a member attempts to join the group with an inconsistent sub-protocol. You would see this if you started up a distributed Kafka connect worker with the default configuration and then tried to use kafka-console-consumer.sh  with the group ID connect-cluster . Rather than overloading this error code, this KIP introduces the new error code INCONSISTENT_GROUP_TYPE  which is similar but applies to the group type (consumer, share, classic) rather than the sub-protocol within the classic consumer group protocol.