One-Second Deploys? We Didn't Believe It Either

One-Second Deploys? We Didn't Believe It Either(blog.railway.com)

3 points by jaredlunde 1 year ago | 3 comments

jaredlunde 1 year ago |

Hey everyone - I wrote the post and the feature. Happy to answer any questions about the implementation and future of functions on Railway.

techpineapple 1 year ago | |

So, AWS allows you to click around in the console to create and schedule lambda functions, relatively quickly; is this blog post comparing apples and oranges?

If I didn't want any of the maturity of using IAAC, I wouldn't have to do the "cloudformation loop" you describe. Similarly couldn't I make railway take longer if I wanted to add IAAC to it's processes as well?

jaredlunde 1 year ago | | |

It's challenging to cover every case when the surface area is as large as AWS. I picked the one that made me want to make this feature for Railway and shared it.

Even still:

- You can't use NPM when writing from the Lambda editor

- Deploying a function to a VPC w/ DB permissions is still a major challenge. If you go with click ops, you will be clicking for >10 minutes if you know what you're doing. If you mess up and want to start over, you will be stuck waiting for the Lambda's networking to teardown, sometimes longer than 20 minutes.

Thus, I think it's a fair enough comparison.