Package com.rabbitmq.stream

Interface MessageHandler.Context

Enclosing interface:

MessageHandler

public static interface MessageHandler.Context

Information about the message.

Method Summary

Modifier and Type	Method	Description
long	committedChunkId()	The ID (offset) of the committed chunk (block of messages) in the stream.
Consumer	consumer()	The consumer that receives the message.
long	offset()	The offset of the message in the stream.
void	processed()	Mark the message as processed, potentially asking for more messages from the broker.
void	storeOffset()	Shortcut to send a store order for the message offset.
String	stream()	The stream the message comes from.
long	timestamp()	The timestamp of the message chunk.

Method Detail

offset

long offset()

The offset of the message in the stream.

Returns:

the offset of the current message

storeOffset

void storeOffset()

Shortcut to send a store order for the message offset.

See Also:

Consumer.store(long)

timestamp

long timestamp()

The timestamp of the message chunk.

Returns:

the timestamp of the message chunk

committedChunkId

long committedChunkId()

The ID (offset) of the committed chunk (block of messages) in the stream.

It is the offset of the first message in the last chunk confirmed by a quorum of the stream cluster members (leader and replicas).

The committed chunk ID is a good indication of what the last offset of a stream can be at a given time. The value can be stale as soon as the application reads it though, as the committed chunk ID for a stream that is published to changes all the time.

This requires RabbitMQ 3.11 or more. The method always returns 0 otherwise.

Returns:

committed chunk ID in this stream

See Also:

StreamStats.committedChunkId()

stream

String stream()

The stream the message comes from.

Returns:

the stream the message comes from

consumer

Consumer consumer()

The consumer that receives the message.

Returns:

the consumer instance

See Also:

Consumer.store(long)

processed

void processed()

Mark the message as processed, potentially asking for more messages from the broker.

The exact behavior depends on the ConsumerFlowStrategy chosen when creating the consumer with ConsumerBuilder.flow().

The call is a no-op for strategies like ConsumerFlowStrategy.creditOnChunkArrival() and ConsumerFlowStrategy.creditOnChunkArrival(int).

Calling this method for each message is mandatory for strategies like ConsumerFlowStrategy.creditWhenHalfMessagesProcessed(), ConsumerFlowStrategy.creditWhenHalfMessagesProcessed(int), and ConsumerFlowStrategy.creditOnProcessedMessageCount(int, double), otherwise the broker may stop sending messages to the consumer.

Applications should make sure to call processed() only once on each context, as this method does not have to be idempotent. What several calls on the same context does depends on the underlying ConsumerFlowStrategy implementation.