public deployments are highly critical environments, due to being tremendously complex also, increasing amount of automation, the risk of the obsolete amount of resource creation is not something wanted.
I recently heard from company X on their almost 50k billing since several months, nobody realised it up until now. This is not just the DevOps issue obviously. Finance, QA and all parts who should be validating should equally embrace this error.
On the other hand, many people complain about the private tooling is time consuming and so forth.
To create seamless integration of available resources and some automated deployment and monitoring mechanism, I've spent on my first OpenStack deployment a fair 6 months.
Many similaritites with public environments additionally you are in charge of all the operations. Although the initial investment is big, you will gain over time having less complexity and no 0.0005 cent/hour calculations will turn to be an avalanche.
All the networking, control and data plane implementations should be done manually, but there is beauty in it.
Using OpenStack, OpenShift Kubernetes is cool but many folks these days I think prefer to choose managed environments like GKE, EKS or AKS.
Considering 2 redundant deployment for HA platforms, one should really take full ownership of the resource management.
![blog-photo](/static/blogs/blog-12/img.png)