AWS Lambda

Posted on August 29, 2016  •  2 minutes  • 271 words

So a few months back, when I read of AWS lambda, I instantly fell in love with it though not having any immediate use with it.

AWS lambda is perfect for those wish to build microservices. There are many benefits of doing this such as independent deployment, security and scability. But I’m not going to talk about this within this post. This post is merely my experience with AWS Lambda.

Enter AWS Lambda: great product, works like magic and requires no dev ops which is a huge plus at where I work. However, there are a few cons:

Testing locally

There is no easy way to do this locally. Looks like there isn’t many alternative out there except maybe lambda-local

Building native packages

I’ve yet to come across a service that build native node.js package for AWS lambda deployment. For now, I have to spin up a EC2 instance, install nodejs and gcc just to build a package, zip it and download it to my local machine in order to prepare for deployment.

Error handling

This isn’t a problem with AWS lambda particular but more like with microservice. However, AWS lambda doesn’t introduce any solution to deal with this either.

As much as I like to use aws lambda in production, i haven’t yet to see any use cases for that at work to put all these in real world test. I’m sure the future of “function as a serivce” is much more and I would love to go into greater details about it in future posts if I have a chance to put it to production. For now though.. Phew