Hello there, I tried the challenges and found it a good way to jump onto serverless and ML services.
So here's a summary blog as some pros and cons tables anyone may find on the challenges.
Who doesn't love to have a website at the lowest rate?
| Pros | Cons |
|---|---|
| Storage cost can be almost insignificant compared to any server only for hosting the files | Must use web framework that supports building a static frontend |
| Free SSL certs | Origin updates can be slow |
Never use the console to create resources and relax
| Pros | Cons |
|---|---|
| Manage resources natively with no cost | Simple environments are straightforward, complex need steep learning curve |
| Cleanup is 100% guaranteed (no costs after deleting) | Some resources will create extra outside the stack, but not this example |
| Automatic rollback to ensure life of the system | Not pure IaC as it's more a configuration file than software, although some functions exist |
The ideal pipeline, were nothing goes wrong, as there are no tests XD. It's a nice intro to CodePipeline
| Pros | Cons |
|---|---|
| No cost for pipeline, as there's no build | No CDN, but could do as the first challenge |
Other than CodePipeline/Build, Lambda can be almost the new Jenkins, just need to have no long running jobs.
| Pros | Cons |
|---|---|
| Nice intro to converting CW metric data | Tutorial doesn't really cover the lambda deployment, but provides the code to get the csv, good prompt into Lambda! |
MLaaS, can't pronounce it but it's fun! A nice intro to understand that a machine learns from history and tries to do its best to predict with complex maths now available to everyone.
| Pros | Cons |
|---|---|
| It's fun to make a machine learn, especially when looking at the results | The tutorial is very low cost, but beware of dangling resources as machines that learn need to be big to be fast (at least for now). |
Anyone can build this app now, amazing!
| Pros | Cons |
|---|---|
| AWS service integration sorts out a lot of hurdles | Needs more CFN handle to create more resources |
| Serverless means it is not far away from production ready in reliability, scale and availability with low cost | App still needs work on user management in case you are really trying to sell it! |
A framework for full stack management of a web app.
| Pros | Cons |
|---|---|
| Can't get quicker to get a live webapp/API and safer than this | Of course like all frameworks will be limited if you want to do fancy things differently |
| Another serverless scalable, etc gem | Remember to avoid costs when free tier time ends |
Getting deeper into AI, a skill is more about experience so you can get a functionality out but practice makes the master.
| Pros | Cons |
|---|---|
| The power of these tools at the palm of your hand (or mouse) is amazing | As the doc says, it's an ongoing work if you're serious about it - "You can (and should) continue to work on your skill, enhancing its features, fixing any problems, and improving the experience for your end users." |
| Need to see videos for part of the tutorial |
Another quick serverless fun intro to ML, if you want to recognise other pictures, just need to continue with the SageMaker challenge
| Pros | Cons |
|---|---|
| Very quick and serverless | Limited use case |
The all time heavyweight champion of automation, made it to the cloud.
| Pros | Cons |
|---|---|
| Tutorial is straightforward | The tutorial exposes you to security as mentioned, but doesn't tell you how to protect against the world |
| Security minimums are decent, and some extras are built in on AWS EC2 so you don't have to worry it burns in 10 minutes of exposure | Need to learn about Jenkins if you haven't yet, to actually have a job running |
| Only an EC2 server to create on AWS, easy cleanup | Challenge mentioned VPC, if you create your own, need more to cleanup but can make it more secure |