Status

Current state: Under discussionAccepted

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

JIRA:

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

...

At the moment, the ListGroups API always returns all groups. There are metrics that track the number of groups in several states but the only way to know the state of a group is to describe it using the DescribeGroups API. In large clusters with hundreds or thousands of groups, it can be hard for a cluster administrator to identify which groups are stable or which groups are dead and can be deleted for example. Basically this requires describing a large number of groups which is inefficiency inefficient in terms of time for the administrator and load for the cluster.

To improve that use case, I propose to allow listing groups by state. For example listing stable groups allows to immediately tell if several applications are healthy if their groups are all returned without needing to describe each group. This would allow building tools and user interfaces that can easily and quickly display informations about live groups.

Public Interfaces

ListGroups API

Update Bump ListGroupsRequest version to 4 and include a new flexible field "statesStatesFilter".

{
  "apiKey": 16,
  "type": "request",
  "name": "ListGroupsRequest",
  // Version 1 and 2 are the same as version 0.
  //
  // Version 3 is the first flexible version.
  //
  // Version KIP-5184 adds the StatesStatesFilter flexible fieldfield (KIP-518).
  "validVersions": "0-34",
  "flexibleVersions": "3+",
  "fields": [
    { "name": "StatesFilter", "type": "[]string", "versions": "4+",
      "about": "The states of the groups we want to list. If empty all groups are returned with their state."
    }
  ]
}

Update Bump ListGroupsResponse version to 4 and include a new field "group_stateGroupState".

{
  "apiKey": 16,
  "type": "response",
  "name": "ListGroupsResponse",
  // Version 1 adds the throttle time.
  //
  // Starting in version 2, on quota violation, brokers send out responses before throttling.
  //
  // Version 3 is the first flexible version.
  //
  // Version 4 adds the GroupState field (KIP-518).
  "validVersions": "0-34",
  "flexibleVersions": "3+",
  "fields": [
    { "name": "ThrottleTimeMs", "type": "int32", "versions": "1+", "ignorable": true,
      "about": "The duration in milliseconds for which the request was throttled due to a quota violation, or zero if the request did not violate any quota." },
    { "name": "ErrorCode", "type": "int16", "versions": "0+",
      "about": "The error code, or 0 if there was no error." },
    { "name": "Groups", "type": "[]ListedGroup", "versions": "0+",
      "about": "Each group in the response.", "fields": [
      { "name": "GroupId", "type": "string", "versions": "0+", "entityType": "groupId",
        "about": "The group ID." },
      { "name": "ProtocolType", "type": "string", "versions": "0+",
        "about": "The group protocol type." },
      { "name": "GroupState", "type": "string", "versions": "4+", "ignorable": true,
        "about": "The group state name." }
    ]}
  ]
}

This KIP bumps the version of this request to avoid any version incompatibility issues with older broker. See rejected alternatives.

If the version is 4 or above, if StatesFilter is If State is not specified or if it's empty in the request , from version 3, all groups will be returned with their states. If StatesFilter is set to a list of group states, only states that are currently in these states will be returned. At the moment, consumer group state is already serialized as a STRING in DescribeGroupsResponse, so serializing it the same way here. 
With this KIP, having Describe authorization on the Cluster resource allows to retrieve the state of all groups. Previously this required having Describe authorization on each Group

When attempting to list groups with states using an older broker, an UnsupportedVersionException will be raised in the client.

AdminClient API

To expose this feature in the AdminClient API, ListConsumerGroupsOptions will be updated.

public class ListConsumerGroupsOptions extends AbstractOptions<ListConsumerGroupsOptions> {

    /**
     * Only groups in these states will be returned by listConsumerGroups()
     * If Bynot defaultset, all groups are returned without their states
     * throw IllegalArgumentException if states is empty
     */
    public voidListConsumerGroupsOptions inStates(List<ConsumerGroupState>Set<ConsumerGroupState> states) { }
 
 
    /**
     * Returns the list of States that are requested
     */
    public List<ConsumerGroupState>Optional<Set<ConsumerGroupState>> states() { }
}

Similarly, the a state field will be exposed in ConsumerGroupListing:

public class ConsumerGroupListing {
    ...
    private final ConsumerGroupState groupState;

    /**
     * Consumer Group Statestate
     */
    public ConsumerGroupStateOptional<ConsumerGroupState> groupStatestate() {
        return groupStatestate;
    }
}

The state will be UNKNOWN empty when ListGroupRequest ListGroupsRequest < 3 is used, for example when connecting to an older broker.

...

To expose this feature in the command line tools, ConsumerGroupCommand tool will be updated to support filtering by group states. A new argument The existing flag "--statesstate" will be updated to allow to specify specifying a list of group states . This new option will only be valid when used with "–list". When specified, the state of each group will be printed next to along the group id.

For example:

# Existing behaviour
./kafka-consumer-groups.sh --bootstrap-server localhost:9092 --list
group1
group2

# Listing all groups with their states
./kafka-consumer-groups.sh --bootstrap-server localhost:9092 --list --states
group1state
GROUP          STATE
group1         stable
group2 empty
         Empty

# Listing groups in the stable state
./kafka-consumer-groups.sh --bootstrap-server localhost:9092 --list --states stable
group1 stablestate Stable
GROUP          STATE
group1         Stable

Compatibility, Deprecation, and Migration Plan

The default behaviour of ListGroups will stay the same and it will still list all groups if no state is specified.

There should be no compatibility issues, nothing to deprecate or migrateWith this KIP, having Describe authorization on the Cluster resource allows to retrieve the state of all groups. Previously this required having Describe authorization on each Group.

Rejected Alternatives

• Allow specifying a filter in ListGroups to retrieve groups by name or by state using wildcard
  
  • Unclear if it's a common use case
• Add flexible fields without bumping the ListGroups version
  • Keeping v3 would allow cases where brokers could support v3 but not the new optional field. In that case, if a client specifies states in its request, brokers would still return the full list of groups without states. This looked confusing. In case broker don't support a feature, a UnsupportedVersionException is preferred.