Gondor.io Django hosting is now publicly available | Hacker News Reader

Gondor.io Django hosting is now publicly available

gondor.io

87 pointsSeanOC29 comments

It's too bad there's not a development plan available. I have quite a bit of AppEngine projects I'd be interested in porting over, but not if 100 requests/day + 500kb SQL (what I would use during the port) would cost me $20/month.

I completely agree. The simplicity that Gondor offers makes the service very appealing, but I'm not going to pay $100/month to host my 5 Django projects that I'm working on concurrently.

Yeah, that's one thing I noticed too. A small free quota on the order of ep.io's (1 WSGI process that maybe gets shut down when not in use + 50-100 MB data + 16 MB Redis) would probably be a good idea.

Gondor looks fantastic. I love the fact that PaaS is picking up and there is more competition. It is such a fantastic idea.

We have been hosting on DotCloud for our django apps and are extremely happy with it. The fact that we can scale up on demand is great. We can add any number of instances by using a simple push command. On top of that the support that DotCloud provides is superb.

All the best to Gondor and DotCloud, I think the pie is big enough for both of them.

This news was posted yesterday: http://news.ycombinator.com/item?id=2955740

This team of people is ridiculously good. They probably wrote half of the Django packages you use.

Anybody know how well this compares to ep.io? As far as I can tell, ep.io is pretty similar, only Epio targets more frameworks than just Django. For better or for worse.

As a small anecdote, before I purchased a VPS I applied to all of the Django-only hosts and visited each one's IRC channel. Gondor was the only one that actually let me try the beta and they were also the only one that had a human on IRC at the time to help with any problems.

Is it conceivable that a service like this could be distilled into Fabric scripts to be run against any hosting provider? I could do without the auto scaling since it would be easy enough to issue the command manually when needed.

Absolutely. After all, we use fabric for some administrative tasks at AppFog. You will also need to write a provisioning layer so you can add, remove and replace hosts as needed. By the time you are finished, you will have your own PaaS.

Yes. I'll be releasing a Django tool this fall (tentatively titled Prometheus), which is essentially a beefed-up django-admin.py startproject, full of best practices for front and back-end development, and sysops stuff, including a fabric script that lets you do this kind of commit to deploy automatically, with any hosting provider or setup (within reason).

Take a look at https://github.com/dziegler/django-create-project & maybe collaborate?

or Pinax :-)

Pinax development's stalled, and it's only about halfway to what I'm hoping to accomplish. They've also made more than a few architectural differences that are pretty incompatible with how I'm doing things. It'll make more sense when you see Prometheus in action.

Pinax development has not stalled at all, it's quite active.

My apologies, it appears I was looking at an outdated fork on GitHub. Off-topic, but I've always found that Google's rankings of GitHub repos to be complete garbage. It rarely shows the original or most active repo, and when it does instead of the link being to the top-level GitHub page for the project, it's some random issue or wiki page from the project. Endrant.

The architectural differences are still a huge concern, and why I never pursued collaboration back when I started it 6 months or so ago. Pinax is great at what it does, but to expand it into what I want a real startproject replacement to be I'd have to tear out way too much. It'd barely resemble Pinax anymore, which would just serve to confuse users.

That project isn't even in the same league of what I'm trying to accomplish. Collaboration wouldn't really make sense in that I'd basically just replace all 10 files he has there. I've already audited approximately 20 different django "startproject" enhancement codebases for the best that they have to offer, to see if I was missing anything useful after the initial development stages.

Good work Gondor team! It is nice to see another quick deploy service launched. It looks the pricing is a little cheaper than Heroku as well, should be an interesting alternative.

Anyone have any experience with using Gondor versus e.g. Linode?

If I'm reading it right, Gondor costs more than twice as much. But that might be worth it to get a server set up very nicely and good Django-specific support.

On top of what Linode offers (which, in Gondor's case is provided by Rackspace Cloud Servers) Gondor is fully managed with an automated provisioning and deployment platform. It's really another layer on top of what a VPS provider is doing.

Good timing. Will probably attract tons of scared App Engine orphans.

I'm afraid to use Gondor. I'm not sure how solid they are yet, whereas Heroku runs Django apps just fine and is definitely here to stay.

>whereas Heroku runs Django apps just fine How?

Heroku's Cedar stack can host just about anything.

There was a discussion of running Django on it not too long ago: http://news.ycombinator.com/item?id=2612943

Take a look at the Cedar stack... it will happily read and install a requirements.txt file.

http://www.askthepony.com/blog/2011/07/getting-django-on-her...

Gondor has been running for ~2 years as the internal hosting platform for Eldarion. A lot has changed (for the better!) over those 2 years, but it is not a brand new system. This is just a signifying of making it usable for anyone and not just Eldarion.

Is there support for WSGI applications in general, like Flask apps?

Not yet, but there's a customer feedback page for it. Anyone interested in this should vote this up.

http://support.gondor.io/forums/111943-general/suggestions/2...

I'd be scared of the Tolkien estate.

Reply on news.ycombinator.com