However, since Ghost is open source, anyone has the option to host it for free. Alternatively, setting up Ghost on platforms like Linode or Digital Ocean would require a minimum monthly cost of $5, based on their lowest plan. This pricing is reasonable considering that Ghost recommends a minimum of 1GB memory for optimal performance.
Why run on AWS?
AWS is running a free trial of their t4g.small instance until December 31, 2023.
Effectively, this means no cost for the remainder of 2023.
However, if this free trial of t4g.small instances is over & above the AWS free tier for EC2, which provides 750 hours of compute? I'm yet to figure that out.
What happens after free tier ends?
The on-demand pricing for t4g.small instance in the
ap-south-1 region as per the aws calculator is 8.18 $ per month. This is more expensive than the price offered by Digital Ocean/Linode for instances of similar capacity.
AWS provides the option of reserved instances for up to 3 years, without having to pay any upfront cost. This brings our monthly cost to as low as 3.43 $ if we choose the Instance Savings Plan.
The price could go further down, but based on my findings, arm instances are cheaper than x86 Instances and also have the added benefit of added performance.
Also, pricing varies by region. So you would have to check and deploy your instance accordingly.
- AWS is cheaper in the long run as a VPS, considering the AWS instances with 2 vCPU and 2GB memory costs around 24% less at 3.80$ per month.