Welcome to the second issue of Last Week in AWS.
It’s tempting to view individual updates and tools in the AWS ecosystem as standalone entitites. This is generally a mistake– any ecosystem is a complex web of interconnected dependencies. I wasn’t expecting to be able to demonstrate this so quickly, but Amazon has given a great example.
Last week I highlighted Cloud Custodian. In typical “AWS released a thing that elegantly solves an issue you spent six weeks working around” fashion, Amazon unveiled a smorgasbord of changes to how they handle resource tagging. This is an unequivocal win for Amazon customers– but in the short term, this deprecated a lot of tooling the AWS community had built for this purpose, as well as portions of last week’s highlighted tool.
Community Contributions
Love My Echo has an interview with Kira Hammond about her useful CloudStatus Alexa skill. What distinguishes this from many Alexa skills is that the CloudStatus is useful. “Alexa, ask CloudStatus about all regions / us-east–2” returns information in realtime (or the AWS status dashboard’s version of it) about AWS’s current status in the given region(s). (Should all of AWS be down, Alexa presumably responds with quiet weeping.) While this is impressive in its own right, let’s not discount that she’s fourteen years old and writing better code than many adult professionals.
On the topic of young people in AWS, AWS Kids had few people fooled. It appeared to be a series of new cloud service offerings aimed at children. It wasn’t hard to see through the joke– while a hilarious spoof, it lacks a realistic AWS name such as “Amazon ClownShoes.”
Meetup’s CTO talks about moving Meetup to the cloud from on-prem datacenters, landing in a new architecture split between AWS and GCP. I’m a sucker for real-world multi-cloud architectures; it’s great to see implementations that transcend individual vendors.
Appaloosa shares a two-part writeup, in Migrating Our Analytics Stack from MongoDB to AWS Redshift. They share the story of their migration from MongoDB to Redshift, highlight the missteps they made, and discuss how they overcame them in an approachable style that the rest of us can learn from.
Choice Cuts From the AWS Blog
Amazon Connect – Customer Contact Center in the Cloud – Amazon launched “Connect,” a call-center-as-a-service offering that enables companies to run and scale call center operations in an elastic way. While simultaneously interesting and valuable, there’s a decent chance that this will become most notable for confusing the living daylights out of people who conflate it with AWS Direct Connect, a physical link between your data center and an AWS region. Adding to the namespace collision, this isn’t the first time something called Amazon Connect has been launched.
New – AWS Resource Tagging API and New – Tag EC2 Instances & EBS Volumes on Creation both speak to enhancements to AWS’s tagging approach. Enforcing specific tags at resource creation time and enhancing how you can interact programmatically with these tags goes a long way towards understanding and filtering elements of your infrastructure. Now when my services start degrading, my monitoring system automatically tags the relevant resources with “Status: Dumpster Fire”.
Amazon Aurora Update – More Cross Region & Cross Account Support, T2.Small DB Instances, Another Region – A series of enhancements to Aurora were released last week. Cross region replication of encrypted databases is a big deal for companies with compliance concerns. You can now share encrypted snapshots of your database with other AWS accounts, making it easier to inadvertently become one of those aforementioned companies with compliance concerns. And in brighter news, t2.small instances help make development environments less screamingly expensive.
Tools
While not a tool in the traditional sense, The Open Guide to Amazon Web Services (along with its attendant slack team) serves as a spectacular starting point to understand the various AWS offerings and their benefits, trade-offs, and gotchas. Community updated and Creative Commons licensed, this serves as a fantastic place to understand how a given AWS offering fits into the larger ecosystem, caveats that people have discovered in its use, and a general overview that collects a lot of those tips that “everyone knows.” Come join the fun!
Is it on AWS both lets you determine if a service is hosted in an AWS environment (with a couple of caveats) and provides a decent framework for how to build similar tools. This is a great example of how to manipulate multiple AWS APIs to solve real problems without needing a computer science background.
Tip of the Week
You may be aware that availability zone mappings aren’t consistent between accounts– in other words, my us-east–1a may be your us-east–1c. This permits Amazon to balance load between AZs (a disproportionate number of people pick ‘a’), and prevents a given AZ from developing a reputation as “the crappy one.”
So what do you do if you need to sync multiple AWS accounts to keep your AZs consistent?
You have two options. The first is to reach out to AWS support. They can sync the mappings for you, but it’s an obnoxious process on their side. If they’re not willing to work with you (read as: you’re not at an enterprise billing tier), you can ferret out which AZ is which by taking a look at the Spot Pricing History. From there “us-west–2b in my prod account maps to us-west–2c in my dev account” is trivial to work out.
But we all know that us-east–1c is totally the crappy one.
…and that’s what happened Last Week in AWS.
