What the best strategy to auto scall a meteor react-native application on AWS

I’ve a meteor react-native application that grow over time, it’s hosted on aws and deployed using meteor-up; as you know, on every client should provide server url or IP with meteor-connect;

For those who has the experience to auto scall it on aws, how do you did it? what’s the best strategy to do it easily ?

Thank’s for your help

The easiest solution is to just use Meteor Cloud.

If you still want to host it “yourself”, you can use Meteor Up Elastic Beanstalk.

1 Like

Thanks you for the quick reply, @mvogt22 ;
Any example to use meteor up elastic beanstalk?
Ps: specially, I use it with meteor react-native, so how to handle the Meteor.connect url in device client ?

@zodern your contribution will be appreciate it, as your are the owner of the package :slight_smile:

I’m building react-native app with Meteor backend too. I use apollo.

have you done the auto scaling, before ?

Just monitor the performance with top and add servers to your load balancer like normal.

If you are able to build an app you can most probably figure out how to edit one text file, and add servers to it. It’s really super easy and normally a task for non-programmers, like an entry level admin job. It’s literally a handful of lines in a conf file.

upstream backend {
    server backend1.example.com;
    server backend2.example.com;
    server backend3.example.com;
    server backend4.example.com;

The manual page is below, always read the manual before you let money leave your wallet.

Of course they will edit my comment again, because I believe he who seeks finds. Merry Christmas everyone! :latin_cross::christmas_tree:

I’m using Google Cloud Compute with it’s auto scaling fearture.
your domain => google load balancer => Group instances. The auto scaling feature can auto add more/remove instances to/from that group.
It works but complicated. Using Meteor Cloud is recommended. It also help Meteor.

Off-topic comments and posts have been edited or deleted.

Please keep on-topic. Thank you.


the purpose is to scale out and down; it’s great idea if I want to use a fix number of server;

I run 5 million daily visitors like this, 800 req/s, 45gbps 95 percentile.

wow, that’s a lots of visitors.

If I used auto scaling and AWS or any of this cloud crap I wouldn’t even break even bud. That’s the truth. You wanna make money, I’m trying to share the truth. There’s a hidden agenda, you need to open your eyes. Servers shouldn’t cost the earth. If you let anyone think for you, you’re paying their mortgage, salary and all of their staffs kids college fee’s. Figure it out.