Going serverless?

Last Update: 10.10.2018. By Jens in Newsletter

Serverless is one of the recent hypes on the block. Is it good? Is it bad?

Probably it depends, like almost anything. There is no black or white. Generic use or don’t use.

Anyways, yesterday I made some free time and got my feet wet. It was a kind of mixed experience.

My test app was a tiny tool I’d want to develop anyways for checking the sales rank of products on Amazon. It already has a sibling written in Python, yep no Java :-), and as they were sharing code anyways I’d also put that new one up in Python.

I used Zappa, which is a tool to help building and deploying Lambda functions written in Python to AWS. Looks promising and actually removed some of the heavy liftings. I basically followed this tutorial. However, where I deviated even before reading the tutorial was, that I was not using my main AWS admin account but started to create one for my Lambda function. Yeah, this is where it turned into a deploy, adjust rights, delete cycle until I finally hade the right permissions set on my user, so Zappa could deploy my whole function. But then, it worked like charm.

So, the first-day learning curve was steep like a cliff. Hopefully, that gets better soon. I have my doubts though. Amazons AWS is complex, sooooo many tools, split docs, etc. No wonder, that companies want you to know that before they hire you…

The interesting question is, is going serverless a must or optional? Does it even make sense for most applications?

That’s the real question.

In my case with two simple tools, both with just a single endpoint and less logic, it totally made sense. It’s tiny, has no dependencies, no business logic and is not part of a bigger system. For regular or complex apps? I am not that sure.

What’s your take on serverless and AWS Lambda?

Want content like this in your inbox each workday irregularly? No BS, spam or tricks... just useful content:

I understand and agree to the privacy policy