Anyone here having experience hosting a Nuxt.js website on a DO droplet with 8GB of RAM and still quite often running out of memory? At least that's what Forge's memory monitoring keeps telling me....
Seems overpowered right? I’m not very experienced with hosting node apps on a VPS - just scaled the droplet up to check if the issues continues. Normally I’d use one with 2GB for a standard laravel / php website and almost never run into any problems.
This droplet however hosts two node apps - a Nuxt.js website with server side rendering and an instance of Strapi CMS. And as I said, the memory consumption is just off the roof…
I'd check the packages used in the project; 9 out of 10 memory leaks usually come from dependencies.
For small traffic
requests per day: Up to 10,000
4GB to 8GB of RAM should be enough.
Medium traffic
requests per day: 10,000 to 100,000
I'd go for 8GB to 16GB of RAM + possibly some scaling
High Traffic
requests per day: 100,000 to 1 million
16GB of RAM and higher, along with load balancing and caching
Of course you should also consider the concurrent users but this is a rough estimate just to give you an idea
2gb server should be easily able to handle 300,000 requests per day. even Redis installed on same server.
Yes obviously it depends on the application behaviour and functionality. There might be so much processing needed, but in general applications if we utilise cache enough, if the content on some pages is static, we render those pages statically
so in my opinion 300,000 requests per day are not much.
if we divide it is 3.4 req per second, but obviously in peak time it will be much more than that, but still 300,000 are not very much.
I mostly work on highly interactive apps an not much static content. As you said it depends on the application behaviour and in your case it looks adequate.
