Building Data-Driven Web Apps with Pyramid and SQLAlchemy Transcripts
Lecture: Deployment overview and topology
0:00 Now that we've built our web app it's time to share it with the world, right? It's great to have a little app that we built but it's basically useless
0:08 if we don't put it on the internet. It is a web app after all. At least put it on a internet for your internal company, right?
0:15 So, that's what this chapter's all about. We're going to see how to deploy our web application onto a standard Linux server in the cloud.
0:24 I want to be clear that this is not the only option there's certainly other ways to put our web app out there. We could go use something like, Heroku
0:32 and just configure Heroku to grab our stuff out of our GitHub repository and launch it into their system. Those to me seem easier
0:41 they're less flexible and often more expensive. But they're easier. So what I want to do is show you how to deploy to a standard Linux server
0:49 run it in some cloud VM somewhere and you can adapt that from DigitalOcean, Linode AWS, Azure, where ever you want to run it.
0:57 So, we're going to do that in this chapter. And that brings us to our overall architecture and topology.
1:04 One of the things we're not going to focus on here is setting up and configuring a database server. I consider that a little bit outside of the scope
1:12 of this course, and you can pick the database server that you want and then configure it so you'll have to fold that in here
1:18 right now, we're just using SQLite, which means as long as the file comes along we have our database.
1:23 So, with that caveat, here's how it's going to work. We're going to go get a box in the cloud which is going to be an Ubuntu server
1:31 probably 18.04, that's what that little icon on the bottom left means. On here we're going to install Nginx.
1:38 Nginx is the thing that people will actually talk to. This listens on port 80 and on port 443
1:45 and for regular HTTP and on 443 for HTTPS encrypted traffic. A request is going to come in here but this does not run our Python code.
1:56 It serves up static files and it delegates to the thing that actually runs our Python code. That thing is called uWSGI uWSGI, I guess it should be.
2:07 Now uWSGI when we run it will handle our Python request. However, we don't want to just run one of them.
2:15 Remember, you may or may not be aware that Python has this thing called the GIL, Global Interpreter Lock
2:20 which inhibits parallelism within a single process. And a lot of the ways people get parallelism in Python is to actually create multiple processes.
2:30 This also has great benefits for fail over or if something goes wrong with some process running one of our requests we can kill it
2:36 and have other processes deal with it. It's not the only way to get parallelism but it's one really nice way. So we're going to do that and have uWSGI
2:43 spin off a whole bunch of itself. This will actually run our Python code. I'm just going to host the Python runtime
2:52 and it's going to launch and initiate a bunch of copies of our website running in parallel. Here we have it configured to run six worker processes.
3:01 So, here's what's going to happen. Request is going to come in hopefully over HTTPS, right, you want to set up some sort of encrypted layer here
3:08 right, that's the way the web's going these days. And once we're inside the server we no longer need encryption so we'll just do a regular HTTP request
3:17 over to UWSGI itself. UWSGI will decide which of it's worker processes is ready to handle it. This time that one is.
3:24 Maybe next time this one's free or maybe that one. It will find one that's not busy or less busy and then pass that request off
3:32 and then return the response back through Nginx back out over HTTPS, to the client. So this is what we're going to set up.
3:40 An Ubuntu server, with these things along with Python 3 and our web app ready to roll.