Consumer prefetch optimization in RabbitMQ - Step-by-Step Execution

Process Flow - Consumer prefetch optimization

Start Consumer

↓

Set Prefetch Count

↓

Receive Messages up to Prefetch

↓

Process Message

↓

Acknowledge Message

↓

Ready for Next Message

↩Back to Receive Messages

The consumer sets how many messages it can handle at once (prefetch). It receives that many messages, processes and acknowledges them, then gets more.

Execution Sample

RabbitMQ

channel.basic_qos(prefetch_count=3)
for method_frame, properties, body in channel.consume(queue):
    process(body)
    channel.basic_ack(method_frame.delivery_tag)

This code sets the consumer to get 3 messages at a time, processes each, then acknowledges them.

Process Table

Step	Prefetch Count	Messages Received	Message Processed	Acknowledgment Sent	Ready for More
1	3	3 messages received	Process message 1	Ack message 1	Yes
2	3	2 messages left from prefetch	Process message 2	Ack message 2	Yes
3	3	1 message left from prefetch	Process message 3	Ack message 3	Yes
4	3	3 new messages received	Process message 4	Ack message 4	Yes
5	3	2 messages left from prefetch	Process message 5	Ack message 5	Yes
6	3	1 message left from prefetch	Process message 6	Ack message 6	Yes
7	3	No more messages or stop	Stop processing	No ack needed	No

💡 No more messages or consumer stopped, so no more messages received.

Status Tracker

Variable	Start	After 1	After 2	After 3	After 4	After 5	After 6	Final
prefetch_count	3	3	3	3	3	3	3	3
messages_in_hand	0	2	1	0	2	1	0	0
messages_processed	0	1	2	3	4	5	6	6
acknowledgments_sent	0	1	2	3	4	5	6	6

Key Moments - 3 Insights

Why does the consumer receive multiple messages at once instead of one by one?

Because the prefetch count is set to 3 (see execution_table step 1), the consumer asks RabbitMQ to send up to 3 messages before acknowledging any. This helps keep the consumer busy without waiting.

What happens if the consumer does not acknowledge a message?

RabbitMQ will not send more messages beyond the prefetch limit until the consumer acknowledges some messages (see execution_table steps 1-3). This prevents overload but can cause delays if acknowledgments are missing.

Why does the number of messages in hand decrease after each processed message?

Because each processed message is acknowledged and removed from the consumer's queue, freeing space for new messages up to the prefetch count (see variable_tracker messages_in_hand).

Visual Quiz - 3 Questions

Look at the execution_table at step 3, how many messages does the consumer still have unprocessed?

A 1

B 2

C 3

D 0

💡 Hint

Check the 'Messages Received' and 'Message Processed' columns at step 3.

At which step does the consumer receive new messages after processing the first batch?

A Step 6

B Step 2

C Step 4

D Step 7

💡 Hint

Look for when 'Messages Received' shows new messages after the initial 3.

If the prefetch count was set to 1 instead of 3, how would the 'Messages Received' column change in the execution_table?

A It would show 3 messages received each step

B It would show 1 message received each step

C It would show 0 messages received

D It would show 6 messages received at once

💡 Hint

Prefetch count controls how many messages are sent at once; see the 'Prefetch Count' column.

Concept Snapshot

Consumer Prefetch Optimization in RabbitMQ:
- Use basic_qos(prefetch_count=N) to limit unacknowledged messages.
- Consumer receives up to N messages before ack.
- Process and ack messages to receive more.
- Helps balance load and avoid overload.
- Adjust prefetch for consumer speed and resource use.

Full Transcript

Consumer prefetch optimization in RabbitMQ means setting how many messages a consumer can get at once before it acknowledges them. The consumer sets a prefetch count, for example 3, so RabbitMQ sends 3 messages. The consumer processes each message and sends an acknowledgment. After acknowledging, RabbitMQ sends more messages up to the prefetch count. This keeps the consumer busy but not overloaded. If the consumer does not acknowledge, RabbitMQ stops sending more messages beyond the prefetch limit. This method balances message flow and resource use efficiently.