I've been working on a few small side projects using modern AI coding tools, and while creating the application itself is quite quick—sometimes just a few hours for a working prototype—I often struggle with the deployment and cloud infrastructure part. I have several questions that typically arise during this phase:
- Which cloud provider should I go with? (AWS, GCP, Azure?)
- What kind of services are best suited for smaller projects (like VMs, serverless options, or containers)?
- How can I design the architecture so it can scale if needed in the future?
- What's the best way to balance cost and CPU for low-traffic projects?
- How can I monitor usage effectively to avoid unexpected costs?
- What's the best way to clean up resources if they've become interdependent?
In some instances, figuring all this out takes longer than building the actual app! I'd love to hear how other developers approach deployment workflows for small projects or minimum viable products (MVPs), whether you configure infrastructure manually every time or utilize tools to streamline the process. Also, if someone has limited DevOps experience, what platform or approach do you recommend to start with?
4 Answers
For simpler setups, I usually start with Docker Compose on a cheap VPS, like Hetzner or DigitalOcean. Caddy simplifies the reverse proxy with automatic HTTPS. Essentially, keep it simple and focus on getting it live quickly, then scale only when necessary. Monitoring is also crucial; I use basic alert systems to catch any issues early.
What do you use for database backups in that setup?
For tailored infrastructures, I prefer using Terraform. It can handle everything from provisioning to cleaning up resources cleanly, but it might be overkill for small projects unless you really want to future-proof your setups.
Interesting perspective! Do you think it’s worth the complexity for MVPs?
I use an affordable local VPS for around €7/month with a Docker Compose setup. Each new project goes into its own Compose stack, and I handle everything through GitHub Actions for CI/CD. It’s super efficient and scales well whenever I need to grow. Caddy for reverse proxy and automatic HTTPS is a game-changer too!
That sounds like such a smooth workflow! How do you manage your persistent data across Docker Compose stacks?
I've learned to steer clear of the big cloud providers for small projects. A cheap VPS is often the way to go. They're cost-effective and straightforward to manage. Just set up Docker Compose on a local VPS, and you're good to go!
What VPS provider do you recommend for getting started?
Great advice! I need to remember not to over-engineer things right at the start.