Junior Message Queues & Async Interview Questions (50 MCQs)

Question 1

What is the main reason to put a message queue between two services?

Accepted Answer

Decoupling producer and consumer so they can work independently

Answer

Making each synchronous HTTP request faster

Answer

Removing the need for a database entirely

Answer

Guaranteeing that messages are never duplicated

Question 2

What does load leveling mean in the context of message queues?

Accepted Answer

The queue absorbs traffic spikes so consumers process at a steady rate

Answer

The queue spreads messages evenly across several databases

Answer

The queue splits large messages into equal-sized chunks

Answer

The queue makes every message the same priority

Question 3

What is the key difference between synchronous and asynchronous service communication?

Accepted Answer

In sync the caller waits for a reply; in async it sends a message and moves on

Answer

Sync works only over HTTP and async only over TCP

Answer

Async always returns the final result faster than sync

Answer

In sync messages are persisted but in async they are not

Question 4

In a messaging system, what is the role of the broker?

Accepted Answer

Receiving, storing and delivering messages from producers to consumers

Answer

Generating messages based on database events

Answer

Running business logic on each message

Answer

Rendering processing results for the end user

Question 5

What is the difference between a queue and a topic in classic messaging?

Accepted Answer

A queue message goes to one consumer; a topic message goes to all subscribers

Answer

A queue accepts only text and a topic only JSON

Answer

A topic stores messages forever while a queue never stores them

Answer

Queues are for internal use, topics only for client communication

Question 6

Which scenario is the best fit for the pub/sub model?

Accepted Answer

An 'order placed' event must reach email, inventory and analytics services

Answer

An image resize job must be done exactly once by a single worker

Answer

A login request needs an immediate response

Answer

Reading a single record from the database by primary key

Question 7

Why is sending a welcome email usually moved to a background job?

Accepted Answer

So signup responds fast and SMTP slowness or failures don't block the user

Answer

Because emails sent in the background get stronger encryption

Answer

Because Node.js cannot send email inside a request handler

Answer

So that emails never end up in the spam folder

Question 8

A user uploads a photo that must be resized into several thumbnails. What's the best approach?

Accepted Answer

Store the file, enqueue a job, and let a worker process it

Answer

Resize every size inside the same request before responding

Answer

Send the raw image to every browser and resize client-side

Answer

Hold the connection open until the CPU becomes free

Question 9

What is BullMQ in the Node.js ecosystem and what does it rely on?

Accepted Answer

A job queue library built on top of Redis

Answer

A standalone broker that replaces Kafka

Answer

An ORM for storing jobs in PostgreSQL

Answer

A built-in Node module for running threads

Question 10

In BullMQ, which call adds a new job to a queue?

Accepted Answer

`await queue.add('send-email', { to })`

Answer

`await worker.push('send-email', { to })`

Answer

`queue.emit('send-email', { to })`

Answer

`await redis.enqueue('send-email', { to })`

Question 11

When does the cron expression `0 2 * * *` run?

Accepted Answer

Every day at 2:00 AM

Answer

Every 2 minutes

Answer

Every 2 hours

Answer

At midnight on the 2nd of every month

Question 12

If a cron job is defined inside a Node app running on 4 instances, what happens?

Accepted Answer

The job runs 4 times unless you use a lock or a central scheduler

Answer

Node automatically runs it on only one instance

Answer

The job never runs because of the conflict

Answer

The work is split so each instance does a quarter

Question 13

What does an acknowledgement (ack) do in a message queue?

Accepted Answer

Tells the broker the message was processed and can be removed

Answer

Tells the producer that a consumer is online

Answer

Encrypts the message so only the consumer can read it

Answer

Raises the message's priority in the queue

Question 14

What is the main risk of auto-ack (acknowledging on delivery)?

Accepted Answer

If the consumer crashes mid-processing, the message is lost

Answer

Messages are always processed twice

Answer

The broker stops sending new messages

Answer

Messages are delivered in reverse order

Question 15

In RabbitMQ, what does `nack` with `requeue: true` do?

Accepted Answer

Puts the message back in the queue for redelivery

Answer

Deletes the message permanently

Answer

Sends the message back to the producer

Answer

Disconnects the consumer from the queue

Question 16

Why are retries important for background jobs?

Accepted Answer

Because many failures are transient, like network timeouts or brief outages

Answer

Because every job always fails on the first attempt

Answer

Because retrying makes a job run faster

Answer

Because without retries a message is never removed from the queue

Question 17

What does exponential backoff mean for retries?

Accepted Answer

The delay between attempts multiplies each time, e.g. 1, 2, 4, 8 seconds

Answer

The number of consumers doubles after each failure

Answer

All attempts happen at a fixed 1-second interval

Answer

The message moves to the front of the queue after each failure

Question 18

Why add jitter (randomness) to exponential backoff?

Accepted Answer

To spread out clients' retries and avoid a synchronized thundering herd

Answer

To randomize message order for better security

Answer

To reduce the total number of retries

Answer

To get messages to the DLQ faster

Question 19

What is a dead-letter queue (DLQ)?

Accepted Answer

A queue for messages that failed processing after retries or were rejected

Answer

A queue for messages the producer hasn't sent yet

Answer

A queue whose messages are deleted immediately

Answer

The main queue with the highest priority

Question 20

What is a poison message?

Accepted Answer

A message that fails every time it's processed, e.g. due to invalid data

Answer

A message containing malware that hacks the broker

Answer

A message whose TTL has expired

Answer

A message that has no consumers

Question 21

What is the competing consumers pattern?

Accepted Answer

Several consumers read one queue and each message goes to only one of them

Answer

Every consumer receives a copy of every message

Answer

Consumers compete for a database lock

Answer

Only the fastest consumer gets messages and the rest shut down

Question 22

In RabbitMQ, where does a producer actually publish messages?

Accepted Answer

To an exchange, which routes messages to queues via bindings

Answer

Directly to the consumer

Answer

To a partition, like Kafka

Answer

To a log file on the consumer's disk

Question 23

How does a `fanout` exchange behave in RabbitMQ?

Accepted Answer

Sends the message to all bound queues, ignoring the routing key

Answer

Sends only to the queue whose routing key matches exactly

Answer

Routes messages based on headers

Answer

Sends each message round-robin to just one queue

Question 24

Which queues does a `direct` exchange route a message to?

Accepted Answer

Queues whose binding key exactly equals the message's routing key

Answer

All bound queues regardless of key

Answer

Queues whose binding matches a wildcard pattern

Answer

Only the queue with the fewest messages

Question 25

In RabbitMQ, what does `durable: true` on a queue guarantee?

Accepted Answer

The queue definition survives a broker restart

Answer

All messages are automatically stored on disk

Answer

Messages never expire

Answer

The queue is replicated across nodes

Question 26

In Kafka, what is a partition?

Accepted Answer

An ordered, append-only log holding part of a topic's messages

Answer

A backup copy of the whole cluster

Answer

A group of consumers sharing a name

Answer

A temporary queue that empties once read

Question 27

What is an offset in Kafka?

Accepted Answer

The sequential position of a message within a partition

Answer

The time between producing and consuming a message

Answer

A globally unique ID for a message across the cluster

Answer

The size of the message in bytes

Question 28

In Kafka, how do consumers in the same consumer group share messages?

Accepted Answer

Each partition is assigned to only one consumer in the group at a time

Answer

Every consumer in the group receives every message

Answer

Messages are spread randomly across consumers

Answer

Only the first consumer in the group gets messages

Question 29

A topic has 3 partitions and your group has 5 consumers. What happens?

Accepted Answer

3 consumers are active and 2 sit idle

Answer

All 5 consumers receive messages

Answer

Kafka automatically creates 2 more partitions

Answer

The group errors and nothing is consumed

Question 30

What's a key difference between Kafka and a traditional queue like RabbitMQ regarding consumed messages?

Accepted Answer

Kafka keeps messages until retention expires, even after consumption

Answer

Kafka deletes a message right after the first read

Answer

Kafka doesn't store messages on disk at all

Answer

Kafka keeps only the latest message per topic

Question 31

How can you reprocess old messages in Kafka?

Accepted Answer

By resetting the consumer group's offset to an earlier point

Answer

By having the producer resend them; there's no other way

Answer

By restarting the broker

Answer

Kafka doesn't allow reprocessing

Question 32

What does `auto.offset.reset=earliest` do for a new Kafka consumer group?

Accepted Answer

With no committed offset, it starts from the oldest available message

Answer

It always restarts from the beginning, even with committed offsets

Answer

It reads only messages produced after it connects

Answer

It auto-commits offsets every second

Question 33

In Kafka, what's the benefit of sending all messages for an order with the same key (`orderId`)?

Accepted Answer

They land in the same partition, preserving their order for that order

Answer

The messages are stored more compactly

Answer

They're spread across all partitions for speed

Answer

Kafka rejects duplicate messages with the same key

Question 34

What is Amazon SQS?

Accepted Answer

AWS's fully managed message queue service

Answer

AWS's managed relational database

Answer

A mobile push notification service

Answer

AWS's managed Kafka offering

Question 35

What is the visibility timeout in SQS?

Accepted Answer

How long a received message stays hidden from other consumers while it's processed

Answer

The maximum time a message is retained in the queue

Answer

How long the producer waits for a send confirmation

Answer

The delay before a newly sent message becomes visible

Question 36

What's the main difference between SQS Standard and FIFO queues?

Accepted Answer

FIFO guarantees order and no duplicates; Standard may duplicate or reorder

Answer

Standard guarantees order but FIFO doesn't

Answer

FIFO stores messages for free

Answer

Standard works in one region and FIFO globally

Question 37

How does the SNS + SQS fan-out pattern work?

Accepted Answer

A message is published to an SNS topic and each subscribed SQS queue gets a copy

Answer

SQS splits messages across several SNS topics

Answer

SNS merges messages from several SQS queues into one

Answer

SNS only encrypts messages and SQS delivers them

Question 38

What does at-most-once delivery mean?

Accepted Answer

Each message is delivered zero or one time; it may be lost but never duplicated

Answer

Each message is delivered at least once and may be duplicated

Answer

Each message is delivered and processed exactly once

Answer

Each message reaches at most one consumer group

Question 39

With at-least-once delivery, what must the consumer be prepared for?

Accepted Answer

Receiving duplicate messages

Answer

Losing messages

Answer

Receiving truncated messages

Answer

The broker never redelivering anything

Question 40

What does it mean for a consumer to be idempotent?

Accepted Answer

Processing the same message multiple times has the same effect as once

Answer

The consumer only ever receives each message once

Answer

The consumer has no state or database

Answer

The consumer processes messages in exact order

Question 41

What's the simplest way to avoid processing a payment message twice?

Accepted Answer

Store a unique message ID (dedup key) and skip ones already seen

Answer

Use auto-ack so the message never comes back

Answer

Increase the number of consumers

Answer

Set the visibility timeout to zero

Question 42

Why isn't processing order guaranteed with multiple consumers on one queue?

Accepted Answer

Because consumers run in parallel and a later message may finish first

Answer

Because the broker deliberately shuffles messages

Answer

Because queues are always LIFO

Answer

Because each consumer reads from the end of the queue

Question 43

In RabbitMQ, what does `channel.prefetch(1)` do?

Accepted Answer

The consumer gets at most one unacknowledged message at a time

Answer

The queue holds only one message

Answer

The consumer reads one message per second

Answer

Only one consumer can connect to the queue

Question 44

In event-driven architecture, what does an 'event' usually represent?

Accepted Answer

A fact that already happened, like `OrderPlaced`

Answer

An instruction to do something, like `SendEmailNow`

Answer

A synchronous request awaiting a response

Answer

A system error that must be logged

Question 45

If queue depth keeps growing steadily, what does it usually indicate?

Accepted Answer

Consumers are slower than the produce rate or have stopped

Answer

The producer is sending messages too slowly

Answer

The system is working ideally

Answer

The broker is compressing messages

Question 46

In BullMQ, how do you add a job that should run 10 minutes later?

Accepted Answer

`queue.add('remind', data, { delay: 10 * 60 * 1000 })`

Answer

`queue.add('remind', data, { timeout: 600 })`

Answer

`setTimeout(() => queue.add('remind', data), 600)`

Answer

`queue.add('remind', data, { priority: 10 })`

Question 47

What is a priority queue used for in a job system?

Accepted Answer

Important jobs (like password reset emails) get processed before low-priority ones

Answer

Jobs are sorted by their size

Answer

Each job runs exactly once

Answer

Older jobs are deleted

Question 48

What messaging model does Redis Streams use?

Accepted Answer

An append-only log with consumer groups, similar to Kafka

Answer

Fire-and-forget pub/sub with no storage

Answer

A plain list queue using only `LPUSH` and `RPOP`

Answer

An exchange with bindings like RabbitMQ

Question 49

What does a TTL (time-to-live) on a queued message do?

Accepted Answer

If not consumed within that time, the message expires and is dropped or dead-lettered

Answer

Hides the message from consumers for that duration

Answer

Sets the maximum processing time for the consumer

Answer

Defines the number of retries

Question 50

To process a 500 MB video through a queue, what's the best approach?

Accepted Answer

Store the file in object storage and send only its URL or ID in the message

Answer

Put the whole file as base64 in the message body

Answer

Split the file into thousands of tiny messages

Answer

Raise the broker's max message size to 1 GB