SciML/DiffEqOnline

Production back-end

Closed this issue · 8 comments

Opening this issue to track progress. I'm working through AWSLambda.jl to see how much of that code we can reuse if we want to go that route.

How would it have to be paid. Do we need to get funding?

Looking at this further I think I'm going to start with Heroku and see if I can get that to work, and I also have a call with JuliaComputing set up for tomorrow morning to hear about their service. I'll forward it to you in case you're available and interested. I can support it through low triple-digit $, and if it looks like it's going to scale beyond that we should start looking around.

whilo commented

Oh, you were faster than me :) I like also to commit the docker files in my repo, so it is really easy to deploy some git repo from itself locally.

Nice, it looks like I'm going to be iced-in this weekend so I should have plenty of time to get this set up.

I just got a base docker image that can use all the needed packages. The article you and @whilo linked was a big help!

I'm splitting it out into a different repo at the moment since some of the services we might use need it to be a standalone git repo. (Still trying to decide between Google, Heroku or AWS Lambda if I can get it to work). If it turns out that it doesn't need to be separate I'll merge it back in.

image

I have the server working from within the container! However, I see a few issues related to this which I will open in the other (possibly temporary) project.

Working. Good job!