SQS & Kafka Interview Quiz

Q: How many questions are in the SQS and Kafka quiz?

There are 30 questions in total: 10 Beginner, 10 Intermediate, and 10 Advanced, randomized on each attempt.

Q: Does it cover AWS SQS specifically or general messaging?

It covers AWS SQS (Standard and FIFO queues) alongside standard Apache Kafka distributed streaming concepts.

Q: Is there any cost?

No, all ToolNotch drills are 100% free and client-side.

Question 1

What is the primary purpose of a message queue in a distributed system?

Accepted Answer

To decouple producers and consumers for async workflows

Answer

To store data permanently in a relational table engine

Answer

To synchronize blocking HTTP calls between microservices

Answer

To provide an in-memory key-value cache layer for nodes

Question 2

In the producer/consumer model, what is the role of a producer?

Accepted Answer

It sends messages to the queue for downstream processing

Answer

It reads and processes messages from the queue

Answer

It manages the queue infrastructure and routing rules

Answer

It monitors queue depth and triggers auto-scaling

Question 3

What is Amazon SQS?

Accepted Answer

A fully managed message queuing service from AWS

Answer

A fully managed relational database service from AWS

Answer

An open-source distributed streaming platform

Answer

A serverless compute engine for batch jobs

Question 4

What is the key difference between an SQS Standard queue and an SQS FIFO queue?

Accepted Answer

FIFO queues guarantee strict order and deduplication; Standard offers best-effort

Answer

Standard queues ensure strict FIFO order; FIFO queues provide random dispatch

Answer

Standard queues enforce payload encryption; FIFO queues do not support KMS keys

Answer

FIFO queues offer cheaper per-message costs; Standard queues enforce hard limits

Question 5

What does 'at-least-once delivery' mean in a messaging system?

Accepted Answer

A message may arrive more than once, requiring idempotent consumer logic

Answer

A message is delivered strictly once with zero duplicate possibilities

Answer

A message is delivered only when downstream consumer workers are online

Answer

A message is retained in queue memory until consumer groups acknowledge

Question 6

What is a Dead Letter Queue (DLQ)?

Accepted Answer

A secondary queue capturing messages that exceed the maximum configured threshold of failed processing attempts

Answer

A high-speed temporary queue that purges messages automatically if they remain unconsumed after sixty seconds

Answer

A critical-priority queue engineered with cross-region replication to broadcast urgent operational outage alerts

Answer

An audit queue that logs forensic checkpoints and metadata for every successfully executed worker transaction

Question 7

What is Apache Kafka?

Accepted Answer

A distributed event stream platform for high-throughput fault-tolerant logs

Answer

A relational database engine optimized for storing tabular time-series

Answer

An in-memory cluster cache built to replace Redis in serverless workloads

Answer

An API gateway designed for routing synchronous REST HTTP microservice calls

Question 8

How does a Kafka topic differ from a traditional message queue?

Accepted Answer

A topic retains an append log allowing multiple consumer groups to read

Answer

A topic drops message payloads immediately after a single consumer read

Answer

A topic restricts producers to a solitary write stream at any given moment

Answer

A topic guarantees sub-millisecond end-to-end transport across all nodes

Question 9

What is a Kafka consumer group?

Accepted Answer

A set of consumers reading jointly from distinct topic partition slices

Answer

A cluster of broker instances that coordinate partition replication sets

Answer

An administrator role that governs topic creation and partition deletion

Answer

A pool of upstream producers that share a unified client identifier tag

Question 10

Why is strict global message ordering difficult to guarantee at scale in distributed messaging systems?

Accepted Answer

Network jitter and parallel node workers turn global ordering into bottlenecks

Answer

Ordering requires relational database storage engines unsupported in message queues

Answer

TCP network layers inherently scramble incoming byte stream frames under load

Answer

Brokers deliberately randomize message delivery order to balance worker consumption

Question 11

What is the SQS visibility timeout and why does it matter?

Accepted Answer

The period during which a received message is hidden from other consumers, preventing duplicate processing

Answer

The maximum time a message can stay in the queue before being deleted automatically

Answer

The delay before a newly sent message becomes visible to consumers

Answer

The time window allowed for a consumer to authenticate before reading a message

Question 12

What is the advantage of SQS long polling over short polling?

Accepted Answer

Hides in-flight messages from other consumers while a worker processes them, reappearing if timeout expires unhandled

Answer

Marks unacknowledged messages as poison pills and discards them immediately to prevent worker starvation faults

Answer

Persists in-flight payloads into encrypted Amazon S3 buckets to guarantee durable retention during host machine crashes

Answer

Restricts queue client connections to a single concurrent worker thread to guarantee strict sequential data consumption

Question 13

What are the default and maximum message retention periods for an SQS queue?

Accepted Answer

Default 4 days, maximum 14 days

Answer

Default 1 day, maximum 7 days

Answer

Default 7 days, maximum 30 days

Answer

Default 1 hour, maximum 24 hours

Question 14

How do Kafka partitions enable parallelism in message consumption?

Accepted Answer

Partitions divide topic streams across consumers for parallel horizontal reads

Answer

Partitions encrypt messages into distinct physical storage volumes on disk

Answer

Partitions compress records to increase worker execution throughput per second

Answer

Partitions buffer producer batches to avoid broker memory starvation faults

Question 15

What is the difference between a committed offset and the latest offset in Kafka?

Accepted Answer

Committed offset is consumer confirmed; latest offset is the partition tip

Answer

Committed offset is next to write; latest offset is the oldest preserved log

Answer

Committed offset resides in SSD blocks; latest offset lives in broker memory

Answer

Committed offset governs producer batches; latest offset governs consumer lag

Question 16

What triggers a Kafka consumer group rebalance?

Accepted Answer

A consumer joins, leaves, or times out, redistributing partition assignments

Answer

A topic administrator dynamically increases partition counts in production

Answer

A producer sends message payloads exceeding maximum configured segment limits

Answer

A broker cluster triggers ZooKeeper or KRaft metadata elections across nodes

Question 17

What is the purpose of a Kafka replication factor, and what does ISR mean?

Accepted Answer

Replication defines copy count; ISR is the set of caught-up partition replicas

Answer

Replication sets queue TTL; ISR measures internal segment retention windows

Answer

Replication sets consumer concurrency; ISR tracks offset lag across brokers

Answer

Replication defines zstd compression; ISR logs checksum indexes for segments

Question 18

Which Kafka configuration achieves exactly-once semantics end-to-end?

Accepted Answer

Enabling idempotent producers with transactional APIs while configuring consumers with `isolation.level=read_committed`

Answer

Setting acks=0 on producers to bypass retries and using single-partition FIFO topics with manual consumer offset commits

Answer

Disabling batching on producers and delegating message deduplication to a zero-latency fault-tolerant ZooKeeper cluster

Answer

Using synchronous replication across all brokers while allowing consumers to auto-commit offsets for every received payload

Question 19

What is the primary difference between Amazon SQS, Amazon SNS, and Amazon EventBridge?

Accepted Answer

SQS is a point-to-point queue for async tasks; SNS is a pub/sub fanout engine; EventBridge is an event bus with schemas

Answer

SQS handles heavy real-time data streaming; SNS processes offline batch jobs; EventBridge connects IoT telemetry nodes

Answer

All three services share identical underlying protocol engines, differing solely in how AWS calculates throughput billing

Answer

SNS persists payloads for fourteen days like SQS; EventBridge operates in memory without content-based JSON filter rules

Question 20

What is Kafka log compaction and when should you use it?

Accepted Answer

A retention policy keeping only the latest value for each distinct record key

Answer

A background daemon compressing log segments with gzip to save disk volume

Answer

A Kafka Streams aggregation merging dual topics into a unified sorted stream

Answer

A cleanup process deleting all messages older than the standard retention TTL

Question 21

What are the three files that make up a Kafka log segment and what does each store?

Accepted Answer

.log stores payload bytes; .index maps offsets to bytes; .timeindex maps timestamps

Answer

.log stores offset metadata; .index stores binary records; .timeindex logs lag

Answer

.log stores group state; .index stores producer tokens; .timeindex tracks replicas

Answer

.log stores compressed batches; .index tracks schema IDs; .timeindex logs TTLs

Question 22

What is a hot partition in Kafka and how does key-based partitioning cause it?

Accepted Answer

A partition receiving excessive writes because a hot key hashes to that partition

Answer

A partition holding excessive replicas; key hashing places replicas unevenly

Answer

A partition temporarily disconnected; key hashing ignores it until node restart

Answer

A partition running compaction; round-robin distribution prevents such compaction

Question 23

How does back-pressure manifest in a Kafka consumer chain and what is a common mitigation strategy?

Accepted Answer

It appears as consumer lag; consumers handle it by scaling instances or tuning poll intervals and max.poll.records

Answer

Back-pressure appears as producer throttling; the mitigation is to increase the number of topics

Answer

Back-pressure appears as broker disk saturation; the mitigation is to enable log compaction

Answer

Back-pressure appears as ZooKeeper timeouts; the mitigation is to upgrade to KRaft mode

Question 24

How does Kafka Streams differ from Apache Flink and Apache Spark Structured Streaming for stateful stream processing?

Accepted Answer

Kafka Streams is an embedded client library, while Flink and Spark require dedicated compute clusters

Answer

Kafka Streams is a batch framework; Flink and Spark are pure streaming engines

Answer

Kafka Streams requires Kubernetes; Flink and Spark run on bare metal only

Answer

Kafka Streams does not support stateful operations; Flink and Spark do

Question 25

What is the throughput limit of an SQS FIFO queue and how do message group IDs affect it?

Accepted Answer

FIFO queues guarantee strict order and deduplication at limited throughput; Standard queues offer nearly unlimited throughput

Answer

FIFO queues support unlimited throughput; message group IDs have no effect on throughput

Answer

FIFO queues support 10,000 messages/second; each group ID adds 1,000 messages/second of capacity

Answer

FIFO queues support up to 1,000 messages/second regardless of group IDs

Question 26

Why is a schema registry important when using Avro or Protobuf with Kafka, and what problem does it solve?

Accepted Answer

It centralizes schemas so producers and consumers validate contracts independently and enforce compatibility

Answer

A schema registry encrypts messages in transit; without it, Avro payloads are sent in plain text

Answer

A schema registry is a Kafka broker plugin that validates message size limits

Answer

A schema registry replaces ZooKeeper for cluster metadata management in newer Kafka versions

Question 27

What metrics and tools are commonly used to monitor Kafka consumer lag in production?

Accepted Answer

Monitor lag via consumer group metrics or Prometheus; scale consumer instances up to the partition limit when lag spikes

Answer

Consumer lag can only be monitored via Kafka's built-in web UI dashboard

Answer

Consumer lag is visible only in AWS CloudWatch when using Amazon MSK

Answer

Consumer lag is tracked by reading the __consumer_offsets topic directly from the application code

Question 28

What is the transactional outbox pattern and how does it solve dual-write problems between a database and a message broker?

Accepted Answer

Writing business state and event records in a single database transaction, then relaying them asynchronously to Kafka

Answer

The outbox pattern stores messages in an SQS FIFO queue before writing to the database, ensuring the database only updates if the queue accepts the message

Answer

The outbox pattern uses two-phase commit across the database and Kafka broker to guarantee simultaneous writes

Answer

The outbox pattern buffers messages in application memory and flushes them to Kafka after the database transaction commits

Question 29

What is Kafka Connect and what are its two main connector types?

Accepted Answer

A pluggable framework connecting Kafka with databases and storage using declarative source and sink connectors

Answer

Kafka Connect is a client library for building custom consumer groups; its two types are polling and push connectors

Answer

Kafka Connect is a monitoring tool; its connector types are metrics and logs connectors

Answer

Kafka Connect is a schema registry extension; its types are Avro connectors and Protobuf connectors

Question 30

How does cross-region replication differ between SQS-based fan-out and Kafka MirrorMaker 2?

Accepted Answer

SNS fans out to multiple regional SQS queues directly, while Kafka uses tools like MirrorMaker 2 for cross-cluster replication

Answer

SQS fan-out and MirrorMaker 2 are functionally identical; the choice depends only on cost

Answer

SQS fan-out replicates messages using Kafka Connect; MirrorMaker 2 uses AWS Data Sync

Answer

MirrorMaker 2 works only with exactly-once semantics; SQS fan-out works only with at-most-once semantics

SQS & Kafka Interview Quiz

What the SQS and Kafka interview quiz covers

Difficulty levels

Question topics covered

How to use the SQS and Kafka quiz

Frequently asked questions about the SQS and Kafka quiz

RabbitMQ Concepts

System Architecture

Database Design

SQS & Kafka Interview Quiz

What the SQS and Kafka interview quiz covers

Difficulty levels

Question topics covered

How to use the SQS and Kafka quiz

Frequently asked questions about the SQS and Kafka quiz

Related Interview Drills

RabbitMQ Concepts

System Architecture

Database Design