Meteor for very small/light websites?


#1

Hi, what do you think about using Meteor for very small/light websites?

Let me tell you why I am asking you that question.

I have pretty good knowledges for Meteor as I am using it for several years now. Before using Meteor I was creating websites using Symfony in PHP. Now I am creating web apps with Meteor and React.

In a few weeks I am going to start web development as a freelancer. To begin, I will target easy developments like small business websites.

I can imagine several options. First, the most standard:

  • Wordpress
  • HTML + CSS only
  • HTML + CSS + React
  • Static website generator like Wintersmith

Or some more powerfull stuff like:

  • Symfony Framework (PHP)
  • Meteor

I don’t really want to use PHP, I want a modern technology and really like javascript, so I was thinking about either “HTML + CSS + React” or Meteor.

If I use pure HTML/CSS/React, then I can use a simple hosting plan and FTP. If I use Meteor, I get a more scalable website for later, but requires more work on servers maintenance…

There are a lot of pro and cons for each solution, that’s why I would like to know what you think about it.
Last question: is it a good idea not to use a CMS for small websites?

Thanks!


#2

You could use CMS if database choice is not a concern.

Between PHP and Javascript. You could consider other languages like Go, RubyonRail, Swift or Rust as a modern language and more like Javascript whereas PHP isn’t and capable of scaling. Don’t take my words, you’ll be better discuss other options in their Slack channel or try out the demos to get the idea.

It’s certainly a good idea to plan ahead and avoid rewrite if it doesn’t meet your requirements e.g. eCommerce, financial, distributed, mobile. If you could use one language to run on all devices and platforms. You are welcome to try it, look for it, I found interesting discussion on one language will supported everywhere and less complicated than Java and PHP.

The incentive to understands the underlying code in libraries and tools that can detect errors in compile time before you publish your code is important, otherwise, like many frameworks, you’ll only know 50% of the working parts, there’ll always be vulnerability in 50% of the code.


#3

I use Meteor for everything, big or small, static or highly data-driven. I don’t know whether it’s a good idea or not, but I don’t have the time or inclination for anything else.

With mup being supported again (or with pm2-meteor, even if mup wasn’t being actively developed), deployment becomes as easy as using FTP (it just takes a little longer for the changes to go live).

What I do know is that I can run half a dozen, low-traffic, mostly-static meteor sites on a $10 D.O. droplet.


#4

Thank you. If I use anything else than PHP, CMS or not, then comes the problem of hosting don’t you think?
I am looking for the cheapest hoster I can find as small websites work fine on those, but the only problem is that there are all PHP only.

For my Meteor App I am used to use a VPS, that I can find for almost the same price as a shared hosting and I use meteor up to deploy.

@babrahams your answer is interesting… I was thinking of using one VPS for each website, even if it only takes 10% of the server ressources, because in that way I can handle invoices more easily. If my client website grows and the server becomes overloaded, then I just need to ask him to buy a higher plan and I can argue about it showing how much visitors it gets every day for exemple. What you do @babrahams is something I should think about, running several websites on one VPS, but how do you handle the case of one server that needs to be upgrade because it gets more traffic? If you tell your client the server is overloaded, don’t you think he will tell it is because he is not the only one using that server?

I’m not sure to be clear, but I’m just not sure how to handle that in a business way. It seems way more complicated that using one VPS for each client.

PS: what hoster are you using @babrahams?


#5

Digital Ocean is the VPS provider I’m using.

If I needed to scale one of the sites, I’d probably move it straight over to Galaxy and pass the cost on to the customer.

As for identifying which site is eating up all the server resources, Kadira is perfect for that. In the absence of Kadira, I guess I’d just have to take a look at the nginx logs and see which app is getting the most traffic.


#6

I hope that Meteor 1.5 is even better for this use case, with the dynamic module loading! I think the initial bundle size should get a lot smaller :]


#7

Just a FYI. If you use a managed database solution (such as mLab) and Galaxy for hosting, you have next to no work on server maintenance.

Prices are slightly higher than other options ($26-28 a month), but they are still very low for any “serious” website (and you can include this in hosting fees for the customer if you go that route - you can say it’s for hosting and maintenance, but realistically you won’t need to do any maintenance).

We first put our business application in to production in November, and have already done over 80 updates. Had absolutely no issues with the simple “deploy” command line, and the server is updated. Every once in awhile we have to throw up a new index on our mLab database if we have some new features that will have a lot of database activity, and that’s it. This would not have been nearly as smooth (and likely not possible) if we were not using Galaxy/mLab.


#8

create-react-app or next.js


#9

Of course Galaxy is the perfect hosting plateform for Meteor apps, but I don’t really see how I can use it in my case. I can’t ask the client to pay $300/year to host his 5 pages website. I’m facing developers that are going to bill $50/year or less for a wordpress hosted on a shared hosting plateform. I don’t think I can convice the client to pay that much more for a simple website.


#10

Thanks, I’ll have a look.


#11

Yeah while Meteor is very convenient to build small sites, you probably don’t want to use Galaxy to host those. However, you can use something like now.sh, which has a Meteor deploy script and I think is cheaper, or at least has a free version!


#12

You can just use your basic Node + Express + React or simple template engine for a simple website. I think Meteor, all that’s going on behind the scenes, is too much for a simple website.

You can get by with the bare minimum on just a simple Express and Node instance.


#13

static sites (serverless) hosted on netlify may be best for that! checkout Gatsby and Phenomic if you want to do React. Cheap to host and nothing is faster.


#14

@dmonni we have developed an Open Source CMS based on Meteor + React - You can give it a try https://github.com/DeligenceTechnologies/Panoplycms - It’s still in Meteor 1.3, going to upgrade it shortly.


#15

Hm… official Meteor partner in India, eh? Too bad MDG didn’t want to give me the time of day when I wrote two Meteor CMS in 3 months. They are still missing the ball.

You must have a better Linked In account, because that’s all they were interested in seeing.


#16

This is a super odd comment and reflects, from the top, the odd position MDG takes in regards to using Meteor for ‘web sites’ while pushing it for ‘mobile [read mostly web sometimes mobile] apps’.

I built my own multi-site Meteor server last year, by myself, and it only took a few months from scratch. I can build it from scratch now in two hours with auto-scaling. Why is MDG so far behind?

Why is Meteor so painful, still, to bundle, deploy and manage? So painful that people often leave Galaxy because their app will no longer deploy there. Sometimes they end up going back, but only because doing it on your own is even more painful, lol.

I truly love Meteor but MDG strategies sometimes leave me quite vexed.


#17

Look, I love Meteor for building interactive applications.

But honestly, for the needs of small business websites there’s nothing Meteor or its half-assed CMS-es will do better than good old Wordpress or Drupal. You even mention you could get away with HTML+CSS only. Then there’s not a single reason to use Meteor/React/…

Your main motivation seems to be you want something more modern. But ask yourself what that gains you for these kind of projects? Modern is mostly a euphemism for immature, incomplete and unstable.

Maybe the right question to ask is if you really want to be building small business websites.


#18

Is it doable? Yes. Is it fast to develop such a project? Hell yes.

But if you want to stick to React and don’t need reactivity, I’d just use Next.js for this project.

Leave Meteor for what it does best.

As for the CMSes thing - I can’t imagine myself using a Meteor CMS.


#19

Hey, thanks for the note, and sorry if you feel like we didn’t have an adequate response. We do look at every project built with Meteor that we see, and we try to reply about any issues people encounter, and work with contributors to make the platform better. However, we also need to focus on moving the core platform forward, and that means we can’t always spend a lot of time helping out projects based on Meteor.

We’re definitely focused on the needs of highly interactive apps, rather than static sites. Meteor can’t be the right tool for every job.


#20

Hi Seba, you’re right, maybe the real question is do I really want to build small websites. Actually, of course I’dd love to build large web app instead of small static website, and this is what I’dd like to do in a few months. But I have to admit that starting as a freelancer for the first time is a little scary haha, that’s why I would like to start it easy.

Now this is why I am thinking using meteor for small website: keep practicing and staying up to date with the Meteor ecosystem. If I only do static websites for the next 6 or 12 months, I’ll have to spend time relearning what I forgot and getting up to date with new stuff that are comming…