Users sometimes ask this question and while it feels easy for us, it looks like magic for someone who is not well-versed into our internals.

Let's add something like this (un-wall of text )

“we read your cloud bill, then take the cost of all the nodes, look for what nodes belong to what cluster, which nodes are worker nodes and which ones are master nodes, then what pods are running on what cluster and namespace and how much CPU, memory, disk, PVCs, etc each one is using, and multiply the cost gathered from the cloud bill by those usage metrics to obtain the $$$ amount that each pod is costing. After that, if the cost model is configured to distribute the cost of the platform and/or the cost of the unallocated capacity, we’ll go and distribute that too. In case you have an on-prem cluster, we'll do the same, but reading from the price list you have created in the RHOCP cost model instead of the cloud bill because there's no cloud bill to read from.”

It'd be nice to add a list of supported costs (CPU, memory, node-attached storage, networking with distribution peculiarities as of this writing, etc) and services (every and any cloud services, including marketplace and private offerings), and which ones we distribute or do not distribute (eg. we don't distribute AWS Route 53, we don't distribute Azure Defender, etc because we cannot attribute it to any workload).