AWS 18

Question 1

What can be problematic with synchronous messaging between applications?

Answer 1

Sudden spikes in traffic can cause issues

Question 2

What is the default time a message can be in a queue? What is the max?

Answer 2

4 days, 14 days

Question 3

What is the max size for messages sent to SQS?

Answer 3

256 kb

Question 4

How can we scale applications processing SQS messages?

Answer 4

We could possibly put the applications processing the SQS on EC2 instances with an ASG;
The ASG will go off of the CloudWatch metric ApproximateNumberOfMessages and set an alarm. The alarm will trigger and ASG will scale.

Question 5

To finish the process of an SQS message, what must we do?

Answer 5

Delete it

Question 6

What is Message Visibility Timeout in SQS?

Answer 6

This means that when a message is polled by a consumer it becomes “invisible” to other consumers. Default time is 30 seconds;
Must be processed in 30 seconds or returned to queue

Question 7

If a consumer needs more time than the default Message Visibility Timeout, what can they do?

Answer 7

Call the ChangeMessageVisibility API to get more time

Question 8

What is Long Polling? Why would we use it?

Answer 8

A consumer can optionally “wait” for messages to arrive if there are none in the queue;
LongPolling decreases the number of API calls made to SQS while increasing the efficiency and latency of your application;

Question 9

What is the SQS Extended Client?

Answer 9

Uses an Amazon s3 bucket as a medium between the producer and consumer. Consumer will consume the message in S3. This helps with large amounts of data.

Question 10

What is the SQS Long Polling API call?

Answer 10

RecieveMessageWaitTimeSeconds

Question 11

What is the difference between FIFO queue and standard queue?

Answer 11

FIFO queue makes sure that first messages in are first messages out. Standard queue does not guarantee order

Question 12

What must be done to the name of the queue if it’s FIFO?

Answer 12

Must have extension .fifo

Question 13

What are the two deduplication methods for SQS FIFO?

Answer 13

Content based deduplication: will do a SHA-256 hash of the message body to make sure same bodies aren’t sent twice
Explicitly provide a Message Duplication ID

Question 14

What are the two deduplication methods for SQS FIFO?

Answer 14

Content based deduplication: will do a SHA-256 hash of the message body
Explicitly provide a Message Duplication ID

Question 15

How would you use the MessageGroupID in SQS FIFO?

Answer 15

you can specify a MessageGroupID so that only one consumer can take that group of messages.

Question 16

What are the protocols for endpoints in subscribing to SNS?

Answer 16

HTTP/S;
Email (JSON);
Amazon SQS;
AWS Lambda

Question 17

What is the SNS + SQS Fan Out method?

Answer 17

A notification is pushed to SNS, where it will be received by all SQS queues. Allows for data persistence, delayed processing and retries of work

Question 18

SNS can send messages to Amazon FIFO queues, true or false?

Answer 18

False

Question 19

We want to create notifications that an object has been created in S3 and we need it to go to multiple services, how do we architect this?

Answer 19

For the combination of event type and prefix in S3, we can only have one S3 event rule. We can send that event notification to an SNS topic and that then published to multiple queues for processing

Question 20

What is Kinesis?

Answer 20

Kinesis provides real time data capabilities from devices to outputs such as Redshift, DynamoDB and S3 bucket.

Question 21

How long does a Kinesis shard keep data?

Answer 21

1 day, up to 7

Question 22

How do we send data to the Kinesis API?

Answer 22

PutRecord API + Partition key that gets hashed (same key goes to same partition)

Question 23

How do we prevent “hot partitions” in AWS Kinesis?

Answer 23

We choose a partition that is highly distributed;
For example, if we have a lot of users, UserID is a good choice but not country_id if the users are mainly in one country

Question 24

What is the Kinesis Client Library?

Answer 24

It is a Java library that helps read record from a Kinesis Streams with distributed applications sharing the read workload.

Question 25

Each shard is to be read by only one KCL instance, true or false?

Answer 25

True

Question 26

How do the KCL’s in Kensis keep track of the workload?

Answer 26

They check their progress into DynamoDB

Question 27

Kinesis records are read in order at the stream level, true or false?

Answer 27

False. They are read in order from the shard level

Question 28

What is Kinesis Data Analytics?

Answer 28

Perform real time analytics using SQL

Question 29

What is Kinesis Firehose?

Answer 29

It is used to load data into Redshift / S3 / ElasticSearch / Splunk

Question 30

We want to scale our SQS FIFO. How can we do this?

Answer 30

Use Group ID and scale the consumers to take care of each group

Question 31

You are preparing for the biggest day of sale of the year, where your traffic will increase by 100x. You have already setup SQS standard queue. What should you do?

Answer 31

Do nothing, SQS scales automatically

Question 32

You would like messages to be processed by SQS consumers only after 5 minutes. What should you do?

Answer 32

Increase the delay queue (DelaySeconds API)

Question 33

Your consumers poll 10 messages at a time and finish processing them in 1 minute. You notice that your messages are processed twice, as other consumers also receive the messages. What should you do?

Answer 33

Increase VisibilityTimeout

Question 34

Your SQS costs are extremely high. Upon closer look, you notice that your consumers are polling SQS too often and getting empty data as a result. What should you do?

Answer 34

Enable long polling

Question 35

You are sending a clickstream for your users navigating your website, all the way to Kinesis. It seems that the users data is not ordered in Kinesis, and the data for one individual user is spread across many shards. How to fix that problem?

Answer 35

You should use a partition key that represents the identity of the user