The most expensive project I had was learning AWS. Their UI/UX is very ... bad at least, and I forgot some resources on different regions, cost me almost 100 dollars.
Side projects are either for learning, so I consider them investments in my career, hobby or to make more money, either way they are worth it. Ofc paying is the last option, I am a big fan of free tiers.
Cloudflare, free GCP/AWS tiers, github pages, Gitlab and others can go for a long run keeping your websites for free, at least for a few hundreds DAU.
Another trick is to get 1 2-4 Core VM and use docker to keep multiple projects there. If a project gets bigger or killed you just move/remove the container.
Their interface is very bad indeed.
Which is why I picked up on Terraform and never looked back
I get to version control my infrastructure and manage it all as code, I can also just run terraform destroy and it all gets removed pretty quickly.
Can't recommend using a code-as-infrastructure tool enough (Serverless is another good one I heard)
Tx! I know about terraform and I recommend it, but it is not good if you want to learn a specific platform.
After you learn it you can use terraform to make it easier, is a framework above cloudformation. I was learning for the certificate, I had to use their UI.
I also used Serverless, after a few manual deploys of my functions, is pretty neat.
Absolutely, free tiers are great. I'm particularly impressed by Gitlab - private repo's and inbuilt CI tools = 👍
Thankfully I've mostly managed to squeeze into the free tiers for everything I've used for side projects on AWS so far. That said, I haven't exactly released any of my (many) side projects for public usage - so would need to keep an eye on the costs at that point...
Yeah, the AWS interface is crazy. I tried out a small project in it once with a free year of hosting. I took everything down, I thought. A month later they told me I owed money for the server that wasn't doing anything 😐
I have done the same thing on Azure. I also ran into an Azure bug that charged me a few grand and took months to clear up.
There are (for GCP anyway, no experience with other serverless providers) pay-as-you-go models for certain components that scale down to 0. That allow you to keep a little-used project running for minimal expense.
E.g., on GCP: App Engine Standard + Datastore.
Yep, my emoji-compress.com is on an appengine like that.
Is not serverless, is just a container and you still have to think of provisioning when u make the config.
Serverless like functions or lambdas are also a good option for event-driven flows.
Personally, I recently launched pointer.nathanheffley.com to little fanfare. It's an extremely simple but reliable Pointing Poker platform. I pay a few bucks per month for hosting, but having a fun, small project to work on that I can use when the need arises makes it worth it.
I host all of my projects on my raspberry pies. I've got 4. I just pay for the domain name.
I've thought about hosting some smaller test projects on a Raspberry Pi I have but I'm too concerned about messing up the security and letting someone gain unauthorized access to my network 😬
I just configured Fail2ban and iptables. I never had a security problem
I run DevLids.com, but it's still small so shared hosting with unlimited traffic works fine. Although I wonder what would happen if it got heavy traffic at some point; if it would just become slow, or if the hoster would ask me to upgrade to another plan.
Other than that I paid for the Kirby-CMS-License and some stickers, so since it launched in June it cost me around 100€.
Hosting the project is usually the only large cost to incur.
Not that I've hit it yet, but I feel that once a service has enough users to exceed a free tier hosting plan, it can usually pay for itself through donations/ads/subscription etc.
I'm the maintainer of EasyFarm which is a free, open source farmbot for the game Final Fantasy XI. I started developing the program in 2012, but was released to the public in 2013, when there was no freely available botting program for the community. The program takes control of your video game character and farms virtual items, so that you can concentrate on the fun things in the game.
There's no way I could charge money for the program because:
I don't even play Final Fantasy XI anymore, but I develop the program, because there are users who still find value in it, but mostly because of my selfish desire to use the program as my personal test bed for new coding techniques.
Development on the program is no longer sustainable and I wish I had tried TDD sooner. Testing the program requires that the game be running and testing all the features takes way too longer. I've started using TDD, but have failed at realizing that dream of being able to test drive a feature and not introducing an AI bug.
Some days, I get really excited about adding big features and other days I feel bad about all the technical debt I've accrued. Many days, I cannot bring myself to even open GitHub, because of all the questions, comments or complaints that are waiting for answers. Occasionally, I get a comment like "Your program has helped me spend less time on the game and more time with my family" and that really raises my spirits. It's an emotional rollercoaster, but I still feel it's worth it.
Wow that's great, it's an amazing feeling when you really, truly help improve somebody's life like that.
$6/month for a Digital Ocean droplet with automated backups, $12/year for a domain. I'm not losing sleep. That should stretch to fit a number of demos before I need to look at upgrading.
We're a place where coders share, stay up-to-date and grow their careers.
We strive for transparency and don't collect excess data.