On-Premises Cost Governance
Introduction
Cost Meter Cards calculate costs for storage and computing resources based on location, vendor, and usage. Virtana Platform enables the application of FinOps principles to on-premises data centres by defining the cost per unit of storage and compute resources. The data can be further segmented based on Business Units, Regions, Resource Vendors, and other relevant categories.
Cost Meter Card selection for cost calculation
Cost calculation utilises Cost Meter Cards, along with cost policies and metric data (from the respective sources, primarily IO), to determine usage costs. To calculate costs for a specific organisation:
Cost Meter Card Selection Criteria: Cost Meter Cards are selected based on the following conditions:
Region and Vendor must not be null, and the charge period must be active (Start Date before now, End Date after now).
Region must not be null, and the charge period must be active.
Vendor must not be null, and the charge period must be active.
Matching and Cost Calculation: The selected Cost Meter Cards are matched with source data based on data centre and manufacturer values. If the keys match, the unit price from the Cost Meter Card is multiplied by the metric value to calculate the cost. The calculated cost is then stored in a new metric in the metric database.
You can also create a Cost Meter Card for a specific entity by enabling the Additional Information fields in Edit Meter Card Fields. See Editing Cost Meter Cards to enable the additional fields.
On-Premise Cost Calculation Process
The on-premise cost calculation process ensures accurate cost tracking and reporting based on selected Meter Cards and their associated data. This process is run as a scheduled job every day. Each job run involves:
Fetching the list of active Cost Meter Cards.
Picking up relevant fields from each Cost Meter Card for matching with the entities, like region, vendor, and charge period.
Fetching the entities matching these fields and their respective usage values.
Calculating cost by multiplying usage values with unit prices and storing results for the day.