Applications Services Summary Flashcards
SQS is an important part of the exam. What is it?
Read the FAQ’s for it
Pull-based message queuing system
Size of SQS messages?
256KB
What’s the visibility timeout for SQS?
Amount of time message is marked as invisible in the queue after a reader picks it up.
Message is deleted from queue if job is processed before timeout expires.
Otherwise message becomes visible again so another reader can process it.
Can result in same message delivered twice
How long can messages stay in the queue?
default is 4 days
can configure from 1 minute to 14 days
What is the visibility timeout maximum?
12 hours
If you need store messages longer than that, don’t use SQS
Does SQS guarantee messages will be processed at least once?
yes
How does long polling differ from short polling?
short polling returns immediately but constantly polls the queue even if queue is empty
Long polling doesn’t return a response until a message is in the queue, or until long poll times out
Long polling can cost less than short polling
What are 2 types of SQS Queues?
Standard and FIFO
Retention Period for SWF vs SQS
SQS - 14 days
SWF - 1 year
API difference of SWF vs SQS
SWF - task-oriented
SQS - message-oriented
How do SWF and SQS deal with duplicated messages?
SWF ensures task only assigned once
SQS can have duplicated messages, you need to handle it
Compare tracking in SWF vs SQS
SWF Tracks all tasks and events
SQS you need to do your own trackingt
3 types of actors
Workflow Starter - application that starts a workflow
Deciders - Controls flow of tasks in workflow execution. If something finishes in workflow, decider chooses what to do next
Activity Workers - carries out activity tasks
Types of SNS Subscribers
http, https email email-JSON SQS Application Lambda function
SNS vs SQS
Both are messaging services in AWS
SNS - push based
SQS - pull based (polls)
What is elastic transcoder?
media transcoder in the cloud
pay based on minutes you transcode and resolution
Core Kinesis Services
Streams
Firehose
Analytics
Streams
data producers send data to streams
streams split into shards
shards pipe data to consumers (EC2 instances)
Firehose
data producers send data to firehose
can query data with lambda
firehose sends data straight to S3
Analytics
Lets you do SQL queries on streams or firehose
