Exposé was engaged to complete a review of this organisation’s development processes and practices for their data team. This review identified extensive manual processes with key dependencies on specific team members to facilitate updates per data product.
Symptomatic of the above processes meant the data team required elevated system access to all products within the organisation. This was flagged as a high-risk issue for the organisation, requiring the data team’s access to be reduced for it to fall in line with other IT team members.
As such, it was requested that the current processes needed to be automated to remove the identified support and security risks that were flagged by the audit.
Exposé worked with the client to define a DataOps framework with the data team, to help them automate their current processes for updating their data products. This required a new software delivery lifecycle (SDLC) to be defined for the data team to use.
With any modern software delivery, the core of the success was moving the team onto a version control system (Continuous Integration). The data team was trained on the best practices when it came to GIT workflows to improve feature integration to products. Additionally helping reduce the risk of quality regression or issues (between team members).
This in turn powered the continuous delivery pipelines, enabling a rigid change control process to be adopted by this client. Helping to ensure all features that were released to test or production are first reviewed, tested, and approved.
The benefits of the new DataOps framework enabled the following:
- Data team elevated access removed, facilitated now by delivery pipelines.
- New defined roles for the data team to facilitate feature delivery.
- More rigid feature delivery:
- Features peer-reviewed by additional team members.
- Features required approval before entering controlled environments (e.g. production).
- Additional automated build processes, ensuring all code is production ready.
- Automated release processes reduced the risk of human error with deployments into production. Also, reducing the outage time required to update production.
- Increased confidence in the change approval board (CAB), resulting in more frequent feature delivery (business value add) over time.
Download our full case study here.