How to deploy the changes¶
Once that the changes into the source code has been finished and the objects necessary to ingest a new entity has been implemented, the next step is to deploy those changes.
Summarizing the changes to deploy:
- Linked services
If the files that will be ingested for the new entity involves a new data source -provider- or a new way to retrieve the data (FTP, DWH, AWS...), it is probably that it has been required to implement new linked services to that provider that must be deployed. Those changes are implemented in the ARM template that defines the Data Factory and the values retrieved from environment data files or from Azure Dev Ops libraries.
- Pipelines and triggers
New pipelines and triggers could also be necessary to move the data from the data sources to the landing zone. Those changes are implemented in new SQL scripts that will be loaded into the databases by the DatabaseBuilder project.
- Custom activities
If any of the new pipelines require a new custom activity, the source code of the activity will be in a new Custom Activity project. That project must be deployed so the assembly of the project will be available for the new custom activity to use. In this case, it will be necessary to update the Build and Release pipelines to take into account the new custom activity project.
Automatic build and release¶
The version control repository selected to work is Git. The source code can be pushed using any Git tool -Visual Studio has a one integrated-. Anyway, when the new code reaches to the dev branch, the Azure DevOps Build pipeline is automatically launched, since it has been configured that way and after a successful build, the Release pipeline is automatically launched.