RabbitMQ Concepts Quiz

RabbitMQ Concepts Quiz

Test your understanding of RabbitMQ exchanges, routing keys, and queue patterns.

1. 1. What is RabbitMQ?

   • A relational database management system
   • An open-source AMQP message broker
   • A REST API gateway
   • A distributed key-value store

2. 2. In a typical RabbitMQ message flow, what is the correct order of components?

   • Consumer → Queue → Exchange → Producer
   • Producer → Queue → Exchange → Consumer
   • Producer → Exchange → Queue → Consumer
   • Exchange → Producer → Queue → Consumer

3. 3. What is a binding in RabbitMQ?

   • A TCP connection between two brokers
   • A rule that links an exchange to a queue (or another exchange)
   • The authentication credential used by a producer
   • A named reference to a virtual host

4. 4. What is the default exchange in RabbitMQ?

   • A fanout exchange named 'default'
   • A pre-declared direct exchange with an empty string name
   • A topic exchange named 'amq.default'
   • A headers exchange that must be declared before use

5. 5. What is the difference between a durable and a transient queue?

   • Durable queues survive broker restarts; transient queues do not
   • Durable queues are faster; transient queues are slower
   • Durable queues can hold more messages than transient queues
   • Durable queues require authentication; transient queues do not

6. 6. What does a consumer send to RabbitMQ to confirm it has successfully processed a message?

   • A NACK
   • A REJECT
   • An ACK (acknowledgement)
   • A CONFIRM

7. 7. What is the difference between basic.nack and basic.reject in RabbitMQ?

   • basic.nack supports bulk rejection of multiple messages; basic.reject handles only one message at a time
   • basic.reject requeues the message; basic.nack discards it permanently
   • basic.nack is only available in AMQP 1.0; basic.reject is AMQP 0-9-1
   • There is no functional difference

8. 8. Which RabbitMQ built-in plugin provides a browser-based management and monitoring dashboard?

   • rabbitmq_shovel
   • rabbitmq_management
   • rabbitmq_federation
   • rabbitmq_stomp

9. 9. What does the prefetch count (basic.qos) setting control in RabbitMQ?

   • The maximum number of queues a channel can subscribe to
   • The maximum number of unacknowledged messages delivered to a consumer at once
   • The maximum message size in bytes
   • The maximum number of connections per virtual host

10. 10. Why is RabbitMQ generally preferred over a relational database for task queue workloads?

    • RabbitMQ stores messages in SQL tables for easy querying
    • RabbitMQ provides built-in push delivery, acknowledgements, and back-pressure without polling
    • A database cannot persist data across restarts
    • RabbitMQ automatically generates SQL schemas for task metadata

11. 11. How does a direct exchange route messages?

    • It broadcasts the message to every bound queue regardless of routing key
    • It routes the message to queues whose binding key exactly matches the message's routing key
    • It uses message header attributes to decide routing
    • It routes based on wildcard patterns in the binding key

12. 12. What is the defining behaviour of a fanout exchange?

    • It routes messages only to queues whose binding key matches the routing key
    • It broadcasts every message to all bound queues, completely ignoring the routing key
    • It routes messages using wildcard patterns (* and #)
    • It sends messages only to queues on the same node

13. 13. In a topic exchange binding key, what does the '#' wildcard match?

    • Exactly one dot-separated word
    • Any single character
    • Zero or more dot-separated words
    • Only numeric characters

14. 14. How does a headers exchange decide which queues receive a message?

    • It matches the message routing key against binding key wildcards
    • It ignores routing keys entirely and routes based on message header attributes matching binding arguments
    • It round-robins messages across all bound queues
    • It routes based on the message body content type

15. 15. What is the conceptual difference between a routing key and a binding key in RabbitMQ?

    • They are identical: both terms refer to the same string set by the producer
    • The routing key is set by the producer on the message; the binding key is set when a queue is bound to an exchange
    • The binding key is set by the producer; the routing key is set on the queue
    • Routing keys are only used by direct exchanges; binding keys are only used by topic exchanges

16. 16. How is a Dead Letter Exchange (DLX) configured for a queue?

    • By setting the 'x-dead-letter-exchange' argument when declaring the queue
    • By publishing directly to an exchange named 'dlx'
    • By enabling the rabbitmq_dlx plugin
    • By setting the 'dead-letter' property on each message individually

17. 17. What is the difference between setting a message TTL per-queue versus per-message?

    • Per-queue TTL applies to all messages in that queue via x-message-ttl; per-message TTL is set in the message expiration property and may vary per message
    • Per-message TTL is always higher priority and overrides any queue TTL
    • Per-queue TTL requires a plugin; per-message TTL is built-in
    • There is no difference: both cause identical broker behaviour

18. 18. What is the key advantage of publisher confirms over AMQP transactions (tx.select/tx.commit) in RabbitMQ?

    • Publisher confirms are synchronous and block the channel until the broker commits
    • Publisher confirms provide asynchronous, lightweight acknowledgement without the heavyweight two-phase commit overhead of transactions
    • AMQP transactions support batch commits; publisher confirms do not
    • Publisher confirms are only available for quorum queues

19. 19. What is the primary purpose of the RabbitMQ Shovel plugin?

    • To mirror queues across cluster nodes for high availability
    • To reliably move or copy messages from a source queue/exchange to a destination queue/exchange, potentially across brokers
    • To expose queues via an HTTP REST API
    • To compress message payloads before storage

20. 20. What is the main architectural difference between quorum queues and classic mirrored queues?

    • Quorum queues use Raft consensus for replication; classic mirrored queues use asynchronous master-mirror replication
    • Classic mirrored queues are faster because they use Raft; quorum queues use two-phase commit
    • Quorum queues can only exist on a single node; mirrored queues span the whole cluster
    • There is no difference: quorum queues are simply renamed mirrored queues

21. 21. How does AMQP 0-9-1 enable multiple logical communication streams over a single TCP connection in RabbitMQ?

    • By opening multiple TCP sockets bundled into one connection object
    • Through channels: lightweight virtual connections multiplexed over a single TCP connection
    • By using HTTP/2 stream multiplexing within the AMQP framing
    • By encoding each logical stream as a separate virtual host

22. 22. What does the AMQP 0-9-1 channel.flow mechanism do?

    • It increases the prefetch window dynamically when the consumer is idle
    • It allows the broker or client to pause/resume message delivery on a channel to apply back-pressure
    • It migrates a channel from one TCP connection to another without interruption
    • It flushes all pending messages from a channel's internal buffer to disk

23. 23. What is the primary operational benefit of declaring a queue as a lazy queue?

    • Lazy queues skip acknowledgement processing to increase throughput
    • Lazy queues write messages directly to disk as they arrive rather than keeping them in memory, reducing RAM usage for deep queues
    • Lazy queues replicate to disk only when the node restarts
    • Lazy queues defer binding evaluation until a consumer connects

24. 24. What is the key architectural difference between the Federation plugin and the Shovel plugin?

    • Federation links are permanent; Shovel links are temporary and delete themselves after a single message
    • Federation provides loosely coupled, broker-to-broker exchange/queue linking with autonomous nodes; Shovel is a simple point-to-point message mover that requires tight coupling
    • Shovel supports AMQP 1.0; Federation only supports AMQP 0-9-1
    • They are functionally identical: Federation is just the newer name for Shovel

25. 25. What does the consistent hash exchange plugin accomplish that a standard direct or topic exchange cannot?

    • It routes messages to exactly one queue chosen by a consistent hash of the routing key, distributing load while keeping messages with the same key on the same queue
    • It replicates messages to all bound queues using a hash-based deduplication mechanism
    • It hashes the message body for integrity verification before routing
    • It ensures round-robin delivery across queues regardless of routing key

26. 26. In a RabbitMQ cluster, what is the functional difference between disc nodes and RAM nodes?

    • Disc nodes persist broker metadata (queues, exchanges, bindings, users) to disk; RAM nodes keep all metadata only in memory for lower latency writes
    • RAM nodes handle message routing; disc nodes handle only persistent message storage
    • Disc nodes can host quorum queues; RAM nodes cannot
    • The terms are deprecated: all modern RabbitMQ nodes are identical

27. 27. How do quorum queues use the Raft consensus algorithm to guarantee data safety?

    • A message is considered confirmed only after a majority (quorum) of replicas have written it, ensuring no data loss even if a minority of nodes fail
    • Raft is used only for leader election; message storage uses a separate two-phase commit protocol
    • Quorum queues write to all replicas synchronously and require unanimous acknowledgement
    • Raft ensures messages are deduplicated across all cluster nodes

28. 28. When both a per-queue policy and a queue declaration argument (e.g., x-message-ttl) are set for the same queue, which takes precedence in RabbitMQ?

    • Policies always win over declaration arguments
    • Declaration arguments always win over policies
    • Declaration arguments (set at queue creation) take precedence over policies for the same key
    • Both are merged additively: neither overrides the other

29. 29. Which pair of RabbitMQ management metrics most directly indicates that consumers are falling behind message producers?

    • publish_rate and deliver_rate: when publish_rate consistently exceeds deliver_rate, messages_ready will grow
    • memory_used and disk_free: a rising memory_used with falling disk_free shows consumer lag
    • connection_count and channel_count: too many channels indicates consumer backlog
    • node_uptime and gc_reclaim_rate: high GC activity confirms consumer lag

30. 30. What is the recommended approach for graceful consumer shutdown in RabbitMQ to avoid message loss?

    • Kill the process immediately: RabbitMQ will automatically requeue any unacknowledged messages
    • Cancel the consumer subscription, finish processing all in-flight messages, send ACKs or NACKs for each, then close the channel and connection
    • Call basic.reject on all messages, then close the TCP connection without sending a connection.close frame
    • Set prefetch to 0 and wait for the broker to drain the queue before closing