Metrics Vault is a Data Vault for the data about the data. Metrics Vault is managed by ODE configuration database. It could be added to the Configuration at any time, and it will start collecting record stats, Vault data load time and record counts. Also Data Vault fields and tables business definition could be recorded in the Metrics Vault.
Technically, your Metrics Vault is just another Data Vault managed by your copy of ODE configuration. All the necessary hubs, satellites and links configuration, also all the stats calculation procedures are get installed with the Metrics Vault.
- Auditing configuration – all the ODE configuration data is audited which is essential for the team development.
- Collecting the data about schedules execution – all ODE executions are audited, including execution time, error messages and record counts. Execution duration is available in ODE core as well, but in Metrics Vault data gets integrated with the configuration, record counts and history of the data.
- Collecting the data stats – attributes get profiled. The history of data profiling is also collected.
- Built-in data dictionary capability – business definitions could be manually entered by users or set up to be uploaded from the data source.
Metrics Vault installation includes writing new records into your configuration database, but it will not have any impact on existing configuration data. To install Metrics Vault, run the installation script from the “Release_scripts” folder.* (Check the installation notes before installation.) This script inserts required configuration, creates two databases, ODE_Metrics_Vault and ODE_Metrics_Stage with all Metrics Vault objects and data collection logic. Databases names are fixed as they are recorded at the configuration data.
Schedule Metrics Vault to run at least daily after the major Data Vault load to pick up changes to the record counts and audit the configuration data. The easiest way to schedule Metrics Vault is SQL Server Agent. Add another step to your Data Vault daily job. Set up the step code:
EXECUTE [dv_scheduler].[dv_process_schedule] 'Metrics_Vault'
Metrics Vault includes a number of views to make metrics data accessible. These views could be a data source for the reporting tool of your choice. Use another script from “Release_scripts” folder to install these views. Note that views should be installed on the same server as the Metrics Vault, otherwise check the references to Metrics Vault in every view.
Among the reporting views, there are few Data Dictionary views. Metrics Vault is not capable to pick up a business definition of the data, but it could store the one you provide. Data Dictionary views combine business definition with some stats and counts, but they also could be used for manual data input. Use SQL update statements on these views to update table or column description. Some code behind these views will not allow direct update of Metrics Vault, but make data flow in the proper direction. Unfortunately, that means that view will not show the change immediately, but only after the Metrics Vault load process.
UPDATE [vw_DD_Columns] SET [ColumnShortDesc] = 'Indicates whether the purchase is made online' , [ColumnLongDesc] = 'Value 1 means that this is an online purchase. 0 means in-store purchase.' WHERE ColumnName = 'OnlineOrderFlag' AND SourceTableName = 'Sales__SalesOrderHeader'
Alternatively, data definition could be loaded to the stage tables from the external source; this needs to be configured in a normal ODE data source way.
* Due to major differences in versions, Metrics Vault only works for ODE instances starting from version 4.
You can download code from the GitHub.