May 29, 2019

Using netlify for a microsite platform?

Stanislav K. @skatkov

I've been using netlify for one of my personal projects and was extremely pleased by the experience. I'm a huge proponent of making things simple and JAMStack was on my radar because of that for a while, netlify in my mind supercharged the concept.

I also maintain a platform to build microsites for Amazon Sellers - https://www.ashop.co . It was heavily inspired my JAMStack, so much that I initially attempted to build headless platform. But while looking for a product market fit I ended up turning my view towards microsites...

Currently, it's a rails monolith, but it was designed in way -- that all microsites are essentially static sites that use API's that platform provides. The problem is -- that I'm a single person maintaining this project, with a huge backlog of features that my clients need.

So, I'm quite limited with my resources -- and can't really dedicate so much time into building all the pipeline required to turning microsites into JAMStack... yeat.. it seems that netlify seems to have everything I need:

  • I need to survive Black Friday, without breaking a sweat or waking up at nights to deal with outrage.
  • Have a build in support for custom domains (with Let's Encrypt support only)
  • Have ability to use my subdomain *.ashop.co with a microsite
  • Be able to easily re-deploy without any downtime every hour or two
  • Have an ability to manage static sites programmatically (through API) -- create sites, changed configuration, re-deploy
  • Have an easy way to distribute microsite across different regions (not only US, but europe and asia)

Netlify seems to have all that and even more.

More I think about this idea, less crazy it sounds... yeat, my mind can't stop wondering about some possible issues I may encounter. I'd appreciate if anyone can give their input. Namely:

  • aShop currently operates at very small profit margin (I plan to introduce premium features to solve that, but so far...). And I have time issue calculating how much it could cost me to maintain...
    So far I didn't pay a cent to netlify for running fbacatalog project. Pricing is based of my usage of lambda/forms... Would that change If I migrate all websites to netlify? is there some sort of traffic limit that those can hit and I have to start paying? Is there any limit/cap on amount of sites I can have on netlify? With my current feature set for every microsite -- basically all site will qualify to be free. Yeat, I feel that it's not a mutually beneficial partnership for netlify...

  • I'm considering applying for accelerator with this project, a lot of those offer free AWS credits. Would it be possible to switch from Netlify CDN to AWS s3?

  • I'm still not sure if I should operate all microsites that my platform houses through my own account on netlify or offer my clients to use netlify on their own? My clients technically are not that 'advanced' and I'm not sure if I can guarantee ease of setup and flawless experience if clients would have to use their own netlify accounts.

  • Maybe someone else has experience of using netlify platform in a similar way and can share some gotchas/positives that they encountered?

Thanks for reading

  1. 3

    It almost feels like you're rebuilding Netlify within Netlify. 😂

    My impression -- and I could very well be wrong -- is that Netlify's margins are made by hosting a good many sites within one VM, and generally adding a margin on top of other AWS services. You'd certainly save a bundle by doing that work and putting it right on AWS yourself, if you're willing/interested in doing so

    1. 2

      For an engineer, it sound exciting to rebuild Netlify -- I bet I would learn a ton. SSL management, how Let's Encrypt Works, how to invalidate cache, having a job that can compile static websites and distribute it over different regions, building my own integration with github..

      Yeat, as a father and a husband, I'm afraid my family will go hungry and homeless. Because they trusted me to do this business thingy.. and I spent so much time doing things that don't matter to my clients.

      Out of this fear, I'm looking for a way to abstract this complexity -- and I see that Netlify can give me this good abstraction

      I do understand, that probably later on rolling my own solution would make ongoing costs lower. Yeat, Spending all that engineering time with amount of customers I have now -- makes this completely unreasonable.

  2. 1

    Very cool idea. It looks like Netlify's API might let you do everything behind the scenes – spin up a site, specify custom domain, add an ssl certificate – so your customers wouldn't even need to know Netlify is involved.

    A couple limits you might hit:

    1. Concurrent builds: With a free plan you're limited to only 1, and to support many concurrent builds it would be $40/each
    2. Bandwidth: 200gb with a free plan – the only time I hit that was hosting video files, but if you're hosting many sites with product images, 200gb is not that much

    I think the big question is how much development time it would take to get things running smoothly. A simpler option might be to just stick with the Rails app and use a sophisticated CDN layer in front for page and asset caching.

    But I freakin' love Netlify...

  3. 1

    I like it. Basically a layer on top of Netlify.

    P.S. I use netlify and JAM stack regularly.