For people who want to make Single Page Applications and responsive websites using Meteor, hosting is a real problem. An average client will request a price comparison between Galaxy vs Godaddy. The answer is going to be a shock for someone wanting to adopt or advocate for this technology.
While Node.js may not be able to run on shared servers, why isn’t that comparison taken into account in scaling hosting services for non-enterprise users?
We are saying that Meteor is easier to learn, has a faster development cycle (read: the more sensible option), but are not really offering an easy, understandable and fair entry point for deployment. That message will be poorly received by some developers and business owners.
As a UX Designer & Hopeful Meteor Developer, scalability goes both ways.
If they do you should explain to them that they can’t really compare a PaaS provider like Galaxy to GoDaddy. They really are 2 different services. You could maybe instead explain to them that something like a Digital Ocean entry level hosting plan is closer to a GoDaddy entry level plan (in terms of cost - both about $5/month), but you can host Meteor apps via Digital Ocean very easily.
I do not know about multiserver deployment, but it was surprisingly easy for me to deploy my hobby project to digitalocean.
I used mupx (mupx not mup. Mup does not work with meteor 1.3). And deployment is almost as easy as galaxy’s.
Also, this guide helped me out alot.
You can host a Meteor app for $5. It’s quite reasonable. The costs of scaling it will largely depend on how you engineer the app. You also have to factor in how much “free” you get with Meteor compared to building something in PHP.
Thank you for the responses! I am looking into Galaxy, Heroku, Modulus, Digital Ocean now. Will I have to de-meteorize? I guess I found cost structure based on usage confusing because it seems open-ended.
Scaling is so often looked at as scaling up. I am looking at scale the other way, for non enterprise, websites for small businesses and individuals, and therefore have to justify cost comparisons. It is easy for performance and UX benefits get kind of lost in this argument.
I didn’t look closely into how this thing works, but it’s must be written for meteor 1.2, so i would advice against that.
Plus, in meteor 1.3 Meteor became more “node-like” and this is what demeteorizer does. So i don’t see demeteorizer being useful.
Is Meteor hosting a price obstacle to its adoption?..An average client will request a price comparison between Galaxy vs Godaddy.
Why target customers who can barely afford these services? How likely is it that someone can afford to pay you $50-100 per hour for software development, yet be unwilling to pay $50-100 per month for hosting? My small business had signed up for a $500 per month ten-container Galaxy plan late last year. Why not work for companies like mine?
Many “average clients” spend more than $50 on bottled water, coffee or printer toner each month. If a few hundred dollars a year for application hosting is prohibitive to your clients, I think you’d be much better off finding new clients.
I think the main issue is that the reason for the difference in pricing is not conveyed properly to the clients/customers/consumers.
The extra price goes into the assurance that everything works without the need for much configuration at all. Businesses have to weigh the pros and cons and whether or not they want to be bothered with the infrastructure side of things. Are they willing to hire someone or delegate the responsibilities of the creation and maintenance of a server(s)?
For hobby, non-enterprises and non-mission critical things, yes a $5 hosting will work and someone that knows how to google can probably set it up as well. But should things go wrong, don’t expect the fastest fix.
What confuses me is that, last time I checked, you need to get Mongo hosting from someone else. Is that still the case? Isn’t that a hassle and addition cost?
Tell them that GoDaddy is like doing your accounting all by yourself in excel. Taking the time to set everything up. You may save $15 a month doing it this way, but it’s much cheaper to leverage something like Quickbooks (Galaxy).
Or some similar metaphor that would click for them.
Or explain that the cost of labor to set it up elsewhere is 3 hours at [insert hourly rate] plus the cost of hosting. Or you can set it up on galaxy for 0 hours labor plus the cots of hosting.
You have to use some VPS even when you are dealing with “classic” PHP.
Frameworks like Symfony2, Laravel also does not work (or it’s painful to run them) on shared hostings.
5$/month reasonable price for hosting
I don’t quite follow - Meteor apps can be deployed anywhere that you can run Node. Seems pretty fair to me!
I think this is more the problem:
From the guide:
But operating Meteor apps correctly, so that your apps work for everyone, can be tricky if you are managing your infrastructure manually. This is why we recommend running production Meteor apps on Galaxy.
Running a single Meteor app instance on different services isn’t a big deal. But since you need proxy server/load balancer with websockets with session affinity, SSL, etc. for anything really worthwhile, the component parts can become expensive if you use AWS tools, or something like Heroku. Each approach seems to have an achilles heel (other than Galaxy).
You’re also tied to MongoDB, and the most popular cloud hosted Mongo starts as ~18/month for 1GB (Yes, I know MongoLab is still free for its’ sandbox tier). You’ll have to pay this even if you only use Mongo for Meteor’s accounts package, as it’s tied to Mongo. You can get up and running with far more storage for less money with much stronger security with AWS RDS, for example.
There’s kind of a space between nothing and full production that’s relatively expensive to work in with Meteor. But these tradeoffs exists for every framework. It’s just important to understand what the tradeoffs are.
i use php, with laravel, and you are right, is dificult convence a client (i from of Argentina the u$s is very expensive for us) whe godaddy, o DonWeb (argentinian hosting) is very cheap, but i try to upload my first project and is hard, and so much expensive,
i try with herouku, aws, gce, and others, finally i like gce (google compute engine) but is very hard to learn, i need 3 weeks to run my aplication and i can write the script to re run automatically meteor. we need a simple solution and cheap for the micro projects (a hotel thre stars with rooms you can select, a restaurant whit a variable menu, etc).
Meteor is great, fast, easy, and powerfull, but the cfs and the deploy is two very hard element to adopt this tecnology. thanks
I haven’t tried it yet, because our apps are self-hosted for the time being.
https://www.digitalocean.com/ provides $20/mo to provide Ubuntu with NodeJS then hook it up with https://mlab.com/ for the MongoDB which can be either $15 or $180 depending on your data set size.
The mlab one is where we’re planning to host our MongoDB installation for our production systems eventually since they would handle the backups and access control for us. We’re still shopping for a Ubuntu/NodeJS provider.
I personally use AWS + Docker. AWS with MupX I also found really easy. Same can be said for DO. Yet to try Galaxy, though I hear very good things!
AWS takes under 2 minutes to spin up a new server for a new app, after the first round of installing everything you need and saving the AMI. Docker makes that even easier.
