Comment by Mark on "Easiest Way to build and Deploy an application in cloud using Django?"

If by cloud you mean some auto-scaling solution, I'm afraid I don't really have experience Django experience in that area.

What I've done in the past is use a VPS, and host each Django website through mod_wsgi, which has the advantage that websites have independent Python interpreters with possibly different versions.

I think if I'd have to do it again, I'd like run it all in Docker instead. Containers for memcached, postgres, nginx, maybe celery, logging, django...

I imagine if it runs in Docker, it should be easy to run on most cloud services, as there are almost no requirements of the host software, except running Docker. But I can't say I've tried.

Thats Mark for this, Ideally I would like to create a platform and the application it's self would be cloned as an instant deploy, but i'll sure take a look at those ones you've mentioned! Much Appreciated! xx

Bridget Sarah Docker would be handy then, you just create more docker containers based on the same image. It's like git: takes a while to get used to, but really handy for coding workflows.

You could go either way with that, do it yourself with a VPS, or use an existing cloud service. I imagine the latter is easier to setup and scale, maybe a bit more expensive. I don't know any examples unfortunately.

Thanks for that, I'll give it ago and lets hope at least dockers a bit more simple to use then aws :D

Bridget Sarah I would not say docker is easier the main advantage is isolation of the userland kernel hence no package collisions and a low abstraction cost on the host system. (so you can have multiple python versions / package in a global space for example) or a service based architecture.

it won't solve the CI problem just the a scale architecture problem and even that is debatable because this gets really ops where most devs forget to use appamor, selinux, kernel spaces. because that is what the 3rd party services are for.

anyhow ... I didn't use words like docker or ansible for the sole purpose to just solve your problem without adding a complex new stack with some projection of future benefits.

just wanted to point out there are implicit problems you can run into without actually solving your main intent. which was a simple system and easy deployment, if I did understand you correct.

Docker does not solve this it solves the orthogonality, isolation, abstraction cost, horizontal scaling problems.

Which is cool if you want to solve this as well go for it :) just don't forget it's a rabbit hole not an easy solution.

Search Hashnode