My question is, are the number of parallel DevOps jobs charged per-second, per-minute, or only per month?
Especially when evaluating setups (we're in the process of migrating from a TeamCity rig), we try different configurations of cloud-hosted vs self-hosted agents, and compare performance vs maintenance work and pricing.
So, say we use 15 cloud-hosted build agents for a week, and then 15 on-prem ones the next week, and then scale back to 2-3 of each for the rest of the month, will we be charged a full month for both 15 on-prem and 15 cloud agents, or do you divide the charge in sub-months, so that we could get charged for 1/4 of a month for each of the 15 on-prem and cloud-hosted agents?
It's difficult to tell from the documentation. Azure Support told me on Twitter to ask in the MSDN forums, here: https://docs.microsoft.com/en-us/answers/questions/263527/per-month-or-per-minute-billing-of-azure-devops-pi.html
But, then I found this answer, which says that DevOps is no longer suppoerted on there: https://social.msdn.microsoft.com/Forums/en-US/72cfd507-06a9-4a43-82d5-58b1eb48df56/azure-devops-pipelines-pricing-and-usage-running-on-deployment-groups?forum=AzureStack
