Has anyone used Webiny before?
It’s open source and looks quite interesting :slight_smile:

1 Like

I think I never had a use case for serverless, but it really seems to be catching on.

1 Like

Tbh I’ve never heard of “serverless”. What is it, really? Only heard of headless but that’s perhaps smth else. Just thought it sounded interesting.

If I understand it, it’s basically you write code and run it, and you pay for the resources or time while it’s running. No managing the stack, just write code and pay to run it.


The “serverless” concept is purely marketing because there is still a server…

But it’s exactly as @Jarland says, you run your code while paying for either the (dedicated) resources or the time it is running on shared resources. Azure Functions for example allow you to trigger code through HTTP, MQTT or other protocols and allow you to compute a few things. It can be helpful for small applications where developers have absolutely no clue on how to manage a VPS. But besides that it’s just very expensive crap. I get paid to create serverless code every so often but I wouldn’t recommend it to anyone.


Thanks @Jarland and @Solaire for the explanation. So not really a need to consider it over a regular VPS in this sense (if you know how to manage it). This reminded me also of Heroku. I think it operates similar to the concept you just explained.

1 Like

Definitely not for personal use. For businesses this is slightly different because they get to sign SLA’s for availability of their application rather than the underlying hardware of course. And while this is much more expensive, it allows them to play the blame game when their application goes offline.

But as you pointed out, Heroku and other application services serve a similar goal except for the fact that serverless applications ought to be very small and scalable. They’re somewhat comparable to microservices but microservices require your own (managed) cluster to run which is more expensive at scale.

An example of a serverless concept would be to receive data over MQTT and having to convert the timestamp before storing it to a database. Assuming you’re using a cloud-based MQTT service and database, creating a full blown application for just this conversion is a bit overkill since it’s literally one line of code, so you’d probably be cheaper off to create a serverless application. But then again, this is only the case for when you’re not running a Kubernetes cluster or rolling your own MQTT / database stack. Cloud environments allow for easy integration of their services (e.g. database, MQTT) with their serverless coding platform.

Marketing talk here


Fair enough. You explained it pretty well, I guess. I wasn’t too sure what to make of it at first. Thanks!

This is a great read on comparing real serverless with current marketing serverless (neo-serverless) written by sqlite developers