I’ve been a netlify user since 2017 and I just deleted all my sites. I can’t risk receiving a $100k bill for toy projects. Your “current policy” is not good enough.
If there was an option to cap billing, or at least some legally binding limit on liability, then I can countenance using netlify.
Until then, it's just not feasible nor worth the risk.
the fact that once it arrives to the limits does not display an error page.
At this point I honestly do not care about they changing their policy, they should have thought that a normal person receiving a 100000$ bill on a free tier shall not been at all on the table in any circumstance, even if they forgive the bill, nobody needs to stress out like that.
I'm not sure about VueJS specifically, but I run everything I can off a $6/m digital ocean droplet (static sites, web apps, git repos, RDBMS, some other custom apps I've written) and it hasn't broken a sweat yet[1].
My understanding used to be that requests will be dropped if my virtual server can't handle it, and I'll have to transfer 10,000TB to get to a $100,000 bill.
In practice, my server will not physically handle the load to serve more than maybe $1000 of data a month; it will fall over before that.
In summary, using a VPS is sorta like an instant hard cap.
[1] Until I tried using Jenkins. Which crashed constantly because apparently 512GB of RAM is too little for what it does. I'm now in the process of writing my own little CD tool that isn't going to go over 30MB of RAM just to run my deployment scripts.
This is the big thing, but I also think that modern out-the-box OSes are pretty damn secure these days.
IOW, the amount of knowledge and time needed to maintain my single VPS is a lot less than the knowledge and time you will need to manage your costs using multiple hosted SaaS suppliers for static hosting, web-app hosting, database hosting, repo hosting, etc.
The only "fix" here is to act like Hetzner and null route upon DDoS, price cap the thing, or offer unlimited bandwidth on the free tier like e.g. Cloudflare Pages.
Uncapped but paid is a recipe for disaster and you'll always be subject to the will of the support staff when something happens. If they can grasp to a straw leading to suspicions that it's not in fact a DDoS attack, you can for example be sure they'll do just that. Just no.
https://www.hetzner.com/dedicated-rootserver/matrix-ax/
With a 48 core Epyc or 80 core arm server, one really shouldnt need much more for a middling project. There are enterprises who run entire services on such hardware.
I did the same last night from my phone. My personal site and a project docs site are just going to not be online for a couple days. Easy choice.
My only hesitation is the 20k file limit; one of our sites is a large Gatsby site that generates certain content programmatically, sometimes using a series of YAML files as the source. Pretty sure we’re over 20k pages by themselves. Been meaning to move this one to Next.js sometime but that’s not on our roadmap anytime soon.
If your web site makes you money in proportion to the amount of views or bandwidth you use, by all means, go with a provider that increases your costs when your traffic rises. But if your web site does not make you money, why not host it somewhere for a flat rate?
I've been on a $25 instance for years for all my sites, which worked as well.