Scale Phase Summary

[00:00:00]
>> Erik Reinert: So what changed? We added state with a bucket. We didn't have any of that before. We also added networking database, cluster and services. We added multiple environments for support. We added staging and production. Outside of that, we also added jobs for running Terraform. We added migrations to run on ecs.

[00:00:18]
You saw us run it directly as well as updated deployments to go on ECS as well. So that deployment script really takes care of the last two part. The new workflow takes care of the jobs. The multiple environment support is all done in our our environment module and Terraform.

[00:00:33]
Inside of our environment module is all of our networking databases, cluster and services. And then everything that we provision exists inside of our state in the bucket. And that's what we've built today. Again, it's so much work just to get one container running. But let's talk about the pros really quickly.

[00:00:51]
So what are the phase pros? We've automated our cloud management pretty much entirely, right? Like we can go from zero all the way to running now with instances and everything else that we need. We consolidated our cloud resources as well. We consolidated it so well that we can create environments based off of it now.

[00:01:11]
So heavy consolidation that made it really easy to manage. We standardized our infrastructure as well. We said, okay, this is what you need to create a network, this is what you need to create a database. This is what you need to create a service. And we standardized that with those modules to make it so that it's really easy to reuse those.

[00:01:28]
But hands down, the biggest one is, yeah, we absolutely made it so that it supports tons of future solutions. I'm sure you guys can look at this and understand the amount of stuff you could build on top of it, how you could grow it out. You can add tons of services to the platform if you want.

[00:01:42]
Now again, you can chop it up, move it around, change it any way that you want because you have such a great control over it. The cons, there are cons and I think you see some of them. For starters, this really does require a team to manage. I would not expect anybody to manage this on their own.

[00:02:02]
It's way too much for one person. It's tons of resources. If it's just you do the app runner approach, do me a favor and go from startup to growth first. And then let's talk about when you actually have users. Going through that process will be a lot more beneficial in the long run.

[00:02:21]
Even though, yes, you could start with all this stuff, it's a lot of complication. Again, I think you could see that now. And that was again, a big part of this course. Requires custom solutions as well. Before in the previous stuff, I didn't have to write any complex kind of weird script or anything like that.

[00:02:41]
This definitely has more complexity with needing to know a little bit more about Cloud and Amazon. And there was a lot of knowledge obviously that I brought in because I was experienced on Amazon and whatnot. But then the last one too is. Requires shared knowledge, right? So again, once you start standardizing things, you start making your own language of like, well, this is a database.

[00:03:03]
Well, that's what that means to us. Or this is a cluster. That's what that means to us. And so like, for example, when we have new hires, sometimes we'll say, yeah, we have clusters. And they'll be like, kubernetes clusters. And we're like, no, ECS shouldn't have assumed. And so yeah, it's a kind of a shared knowledge as you start taking on this management more and more that will be shared amongst your teammates.