Show HN: Sick of running random cron jobs, I built a small Keepalive for my FaaS

stayingalive.vessels.tech

12 points·lwilld··8 comments

If you need to keep your function "warm," why are you using FaaS in the first place? This almost seems like serverless satire!

I don’t understand this either. What’s the point of going serverless with something like Lambda if you’re treating it like a regular old EC2 instance with more overhead? Is it cheaper or something?

I'm pretty sure the "point" is "I drank the koolaid. Turns out I have diabetes and can't have that much sugar but fuck it insulin injections work so pass the koolaid bro"

I'm afraid I don't understand the purpose of this. Are there servers that die if you don't hit their URLs periodically? I use uptimerobot for my servers but it serves a different purpose.

For FaaS (like Lambda), if a function hasn't been hit in a while, it'll be cold started when it is needed next. (Even though it's "serverless" behind the scenes they're likely containers that stay up for a little while to handle subsequent requests; a cold start means there's overhead of starting up the container)

A hack to keep your function warm is to call it via cron or a service like this.

404 and invalid cert :-/

I get it. Well done!

I had the same problem when I first launched https://dndemail.com. Traffic was low and if no one uses the service for an hour or so, the first load in the app was slow.

I built a once a minute Cron to keep the app warm and initial requests quick.

Now I have enough traffic that I don't worry about the problem.

Why once a minute?

Reply on news.ycombinator.com