Gitlab is talking about heading into the private cloud after successfully building a cloud-ready application. The savings are substantial for a small, technology-rich company:
The cloud hosting for GitLab.com excluding GitLab CI is currently costing us about $200k per month. The capital needed for going to metal would be less than we pay for 1 quarter of hosting. The hosting facility costs look to be less than $10k per month. If you spread the capital costs over 2.5 years (10 quarters) it is 10x cheaper to host your own. (My emphasis)
This sounds about right but I don’t think this factors in head count for operating the physical infrastructure. Lets say that two extra FTEs at $15K per month are required, this still one third the cost of AWS. The reaility is $2.4MM is a substantial yearly budget for IT Infrastructure and for an application that already cloud-ready it would go a very long way
For a small company that is focussed on technology adding more headcount is good for capacity. In a team of ten people, adding 2 headcount increases diversity of thinking, ideas and approaches and can be important to spreading out the workload e.g. on call rotation is much improved with 12 people in the rota.
Factors that I thought significant:
- Application fits into a single rack (its a small company)
- Company is focussed on a single customer solution which they own end-to-end.
- Strong technical leadership has insights in the problems of owning infrastructure and is thinking ahead on minimising that.
- Making some good decisions to keep the physical infrastructure simple, plain and easy to handle.
I’ve said many times, public cloud isn’t cheap. This doesn’t make it the right or wrong solution but cheap is not a reason to go to the cloud when operating at any sort of scale. (Mind you, a startup with less then five people will always be better off using public cloud). There isn’t a right answer just use cases for when it works and doesn’t.