Status

Current state: Vote in progress Accepted

Discussion thread: here

Vote thread: here

...

The PATCH /connectors/{connector}/offsets  endpoint will be useful for altering the offsets of stopped connectors. Requests will be rejected if the connector does not exist on the cluster (based on the config topic), if a rebalance is pending, if the connector is not in the STOPPED  state (described below), or if it does not have an empty set of tasks in the config topic. If successful, the request will be met with a 204 ("no content") 200 response and an empty response the appropriate body. The request will fail with a 404 response if the connector does not exist on the cluster, and will fail with a 400 response if the connector does exist but is not in the correct state.

The connector's alterOffsets  method (described below) will be invoked, before modifying its consumer group offsets or source offsets.

Requests to this endpoint will only augment any existing offsets for the connector; they will not implicitly remove any offsets for the connector that are not included in the request body.

All offset alter requests will be forwarded to the leader of the cluster.

Resetting offsets

The DELETE /connectors/connector/offsets  endpoint will be useful for resetting the offsets of stopped connectors. Requests will be rejected if the connector does not exist on the cluster (based on the config topic), if a rebalance is pending, if the connector is not in the STOPPED  state (described below), or if it does not have an empty set of tasks in the config topic. If successful, the request will be met with a 204 ("no content") response and an empty response body. The request will fail with a 404 response if the connector does not exist on the cluster, and will fail with a 400 response if the connector does exist but is not in the correct state. This endpoint will be idempotent; multiple consecutive requests to reset offsets for the same connector with no new offsets produced in between those requests will all result in a 204 response (if they are successful).

The connector's alterOffsets  method (described below) will be invoked, before deleting its consumer group or source offsets. The offsets  argument will contain keys for every known topic partition in the consumer group (if the connector is a sink) or every known source partition (if the connector is a source), each with a null value.

A source offset will only be considered successfully deleted if the Connect worker is able to emit a tombstone to the offsets topic for its partition, and then read to the end of the offsets topic. A request to reset offsets for a source connector will only be considered successful if the worker is able to delete all known offsets for that connector, on both the worker's global offsets topic and (if one is used) the connector's dedicated offsets topic.

If exactly-once source support is enabled on a worker that receives a request to reset offsets for a source connector, it will fence out all previously-running tasks for the connector (if any exist, based on the presence of a task count record in the config topic) before invoking alterOffsets  and resetting the offsets for the connector. Offsets will be reset transactionally for each topic that they exist in: a single transaction will be used to emit all tombstone records for the connector's dedicated offsets topic (if one is used) and another transaction will be used to emit all tombstone records for the worker's global offsets topic.

Requests to reset sink connector offsets will be satisfied by deleting the consumer group for that sink connector (as opposed to deleting all known offsets for that consumer group). Requests to reset sink connector offsets will fail if there are any active members of the sink connector's consumer group.

All offset reset requests will be forwarded to the leader of the cluster.

New target state: STOPPED

Background: target states today

Kafka Connect currently supports two "target states" for a connector: RUNNING  (the default), and PAUSED. The target state for a connector can be controlled using the REST API, using the PUT /connectors/connector/resume  endpoint for the RUNNING  state and the PUT /connectors/connector/pause  endpoint for the PAUSED  state.

When a connector is paused, its tasks continue to exist on the cluster. Many resources allocated for them, including Kafka and Kafka Connect resources such as clients, converters, and transformations, and connector-specific resources such as database connections, file descriptors, memory buffers, remain allocated. This can lead to confusing and sometimes suboptimal behavior when users pause connectors but notice that resources (especially database connections) are still allocated.

If exactly-once source support is enabled on a worker that receives a request to alter offsets for a source connector, it will fence out all previously-running tasks for the connector (if any exist, based on the presence of a task count record in the config topic) before invoking alterOffsets  and altering the offsets for the connector. Offsets will be altered transactionally for the connector's primary offsets topic (i.e., the one that is written to transactionally by source tasks when the connector is running). The transactional ID used for this operation will be ${groupId}-${connector}, where ${groupId} is the group ID of the Connect cluster and ${connector} is the name of the connector.

Requests to this endpoint will only augment any existing offsets for the connector; they will not implicitly remove any offsets for the connector that are not included in the request body.

All offset alter requests will be forwarded to the leader of the cluster.

Resetting offsets

The DELETE /connectors/connector/offsets  endpoint will be useful for resetting the offsets of stopped connectors. Requests will be rejected if the connector does not exist on the cluster (based on the config topic), if a rebalance is pending, if the connector is not in the STOPPED  state (described below), or if it does not have an empty set of tasks in the config topic. If successful, the request will be met with a 200 response and the appropriate body. The request will fail with a 404 response if the connector does not exist on the cluster, and will fail with a 400 response if the connector does exist but is not in the correct state. This endpoint will be idempotent; multiple consecutive requests to reset offsets for the same connector with no new offsets produced in between those requests will all result in a 200 response (if they are successful).

The connector's alterOffsets  method (described below) will be invoked, before deleting its consumer group or source offsets. The offsets  argument will contain keys for every known topic partition in the consumer group (if the connector is a sink) or every known source partition (if the connector is a source), each with a null value.

A source offset will only be considered successfully deleted if the Connect worker is able to emit a tombstone to the offsets topic for its partition, and then read to the end of the offsets topic. A request to reset offsets for a source connector will only be considered successful if the worker is able to delete all known offsets for that connector, on both the worker's global offsets topic and (if one is used) the connector's dedicated offsets topic.

If exactly-once source support is enabled on a worker that receives a request to reset offsets for a source connector, it will fence out all previously-running tasks for the connector (if any exist, based on the presence of a task count record in the config topic) before invoking alterOffsets  and resetting the offsets for the connector. Offsets will be reset transactionally for the connector's primary offsets topic (i.e., the one that is written to transactionally by source tasks when the connector is running). The transactional ID for this operation will be ${groupId}-${connector}, where ${groupId} is the group ID of the Connect cluster and ${connector} is the name of the connector. If this topic is different from the worker's global offsets topic, the connector's offsets will be removed from that topic as well, but without the use of a transaction.

Requests to reset sink connector offsets will be satisfied by deleting the consumer group for that sink connector (as opposed to deleting all known offsets for that consumer group). Requests to reset sink connector offsets will fail if there are any active members of the sink connector's consumer group.

All offset reset requests will be forwarded to the leader of the cluster.

New target state: STOPPED

Background: target states today

Kafka Connect currently supports two "target states" for a connector: RUNNING  (the default), and PAUSED. The target state for a connector can be controlled using the REST API, using the PUT /connectors/connector/resume  endpoint for the RUNNING  state and the PUT /connectors/connector/pause  endpoint for the PAUSED  state.

When a connector is paused, its tasks continue to exist on the cluster. Many resources allocated for them, including Kafka and Kafka Connect resources such as clients, converters, and transformations, and connector-specific resources such as database connections, file descriptors, memory buffers, remain allocated. This can lead to confusing and sometimes suboptimal behavior when users pause connectors but notice that resources (especially database connections) are still allocated.

In addition, connectors and tasks are allocated across the cluster without regard to their target state during rebalance. This can lead to some skew in resource usage (such as network traffic and CPU utilization) across the cluster; in the worst case, all paused tasks are allocated to one set of workers, and all running tasks are allocated to a disjoint set of workers.

...

public abstract class SinkConnector extends Connector {

    /**
     * Invoked when users request to manually alter/reset the offsets for this connector via the REST
     * API. Connectors that manage offsets externally can propagate offset changes to their external
     * system in this method. Connectors may also validate these offsets if, for example, an offset
     * is out of range for what can be feasibly written to the external system.
     * <p/>
     * Connectors that do not manage offsets externally or require custom offset validation need not
     * implement this method beyond simply returning {@code true}.
     * <p/>
     * User requests to alter/reset offsets will be handled by the Connect runtime and will be reflected
     * in the offsets for this connector's consumer group.
     * <p/>
     * Similar to {@link #validate(Map) validate}, this method may be called by the runtime before the
     * {@link #start(Map) start} method is invoked.
     *
     * @param connectorConfig the configuration of the connector
     * @param offsets a map from topic partition to offset, containing the offsets that the
     *                user has requested to alter/reset. For any topic partitions that are being reset instead
     *                of altered, their corresponding value in the map will be {@code null}.
     * @return whether this method has been overridden by the connector; the default implementation returns
     * {@code false}, and all other implementations (that do not unconditionally throw exceptions) should return
     * {@code true}
     * @throws UnsupportedOperationException if it is impossible to alter/reset the offsets for this connector
     * @throws org.apache.kafka.connect.errors.ConnectException if the offsets for this connector cannot be
     * reset for any other reason (for example, they have failed custom validation logic specific to this connector)to this connector)
     */
    public boolean alterOffsets(Map<String, String> connectorConfig, Map<TopicPartition, Long> offsets) {
     */
   return false;
  public boolean alterOffsets(Map<String, String> connectorConfig, Map<TopicPartition, Long> offsets) {
        return false;
    }

}

Compatibility, Deprecation, and Migration Plan

Offsets endpoints

 }

}

Compatibility, Deprecation, and Migration Plan

Offsets endpoints

This feature is fully backwards-compatible with existing Kafka Connect releases. Migration should occur automatically whenever a Kafka Connect cluster is upgraded to a version that supports this feature.

Additional connector ACLs

If exactly-once source support is enabled on a worker, in order to handle requests to alter or reset offsets, the connector's principal must be able to use a transactional ID of ${groupId}-${connector}, where ${groupId} is the group ID of the Connect cluster and ${connector} is the name of the connectorThis feature is fully backwards-compatible with existing Kafka Connect releases. Migration should occur automatically whenever a Kafka Connect cluster is upgraded to a version that supports this feature.

STOPPED target state

Diligent readers will note that the addition of a new target state creates problems with cluster downgrades and rolling upgrades. If a connector is stopped, and a worker running an older version of Kafka Connect either joins or currently exists in the cluster, that worker may not know how to handle the new record in the config topic that includes the request to stop the connector.

...

Scope: The new PATCH /connectors/{connector}/offsets  endpoint

Dependencies: The  The STOPPED state feature and Resetting offsets (internal logic) features

Resetting offsets (internal logic)

...

Dependencies: The STOPPED state and Resetting offsets (internal logic) features

Test Plan

Integration Unit, integration and/or system tests will be added for these cases:

...

• Can stop a running connector
• Can stop a paused connector
• Can stop a stopped connector (i.e., stopping an already-stopped connector is idempotent)
• Can delete a stopped connector
• Cannot see task configuration or status information in the REST API for a stopped connector
• Can resume a stopped connector
• Can pause a stopped connector
• Stopping a failed connector updates its state to STOPPED  in the REST API
• Stopping a connector that fails during shutdown after receiving a stop request updates its state to STOPPED  in the REST API
• Can resume a connector after its Connector  has failed both before and during shutdown after receiving a stop request

...