A global food manufacturer was growing, and operational and analytical reporting demands were increasing, as was the need to support users in different time zones. They were with inconsistent manual build and deploy processes that were slow and error-prone. A strategic DevOps initiative reviewed existing processes, landscape, and tooling and identified the ideal scene—common processes were vital for their delivery lifecycle. After an RFP process and great success with FlexDeploy in another business area, the company brought FlexDeploy on board for the Business Intelligence team.
Challenge
Extensive growth through acquisitions, global expansion, and implementation increased the workload of the Oracle Business Intelligence (OBIEE) team. The challenges they faced prevented them from moving at the speed required to keep up with the demands of the business.
In addition to manual build and deploy processes that increased project development duration and cost and led to system outages, people-dependent processes led to bottlenecks. Code movement responsibilities fell on a small group of administrators, causing project teams to wait for an administrator’s availability to move code and bounce servers. Once code was moved, there was also a wait to address missed or incorrect objects. OBIEE data source names, for example, needed to be changed manually after the repository or web catalog object was moved, leading to even more project development delays.
Custom-built scripts were also people-dependent, requiring staff hours to build, test, and maintain. Custom scripts were built to move database objects through disparate environments, but they often weren’t maintained during testing. This led to missing changes or incorrect table column definitions, resulting in additional project development time needed to create and maintain these scripts, and even more time was needed to run these scripts when project teams had to move to the next environment. With so many manual processes to move code objects, implementation planning became a web of scripts, steps, and resources requiring coordination across groups.
Ultimately, the team needed to remove variation and speed up their deployment and release activities while maintaining and improving the quality of work—essential resources needed to work on more value-added activities rather than administrative tasks.
Solution
The company considered IBM/Urbancode, CloudBees, and Jenkins, but after a successful relationship with FlexDeploy in another business area, they chose FlexDeploy. With FlexDeploy implemented, the team saw improvements across three key areas—faster development, increased control and empowerment, and improved time to production.
Automation replaced manual and scripted tasks, eliminating the need for administrators to do manual moves and server tasks—drastically reducing the time spent creating, testing, and maintaining DDL scripts. Automated, tested, and repeatable processes also reduced outage time and the time to investigate and fix. FlexDeploy also allowed the developer to move code when needed versus waiting for administrators. The team configured FlexDeploy so all moves to production required approvals, enforcing a framework that supported their software development lifecycle strategy.
What’s more, FlexDeploy’s preview option cut time previously spent on validating large DDL scrips. Instead, the team could deploy the DDL changes in preview mode for validation before go-live. They found that implementation planning for releases was simpler as all builds were centralized in one tool. As a result, code discrepancies became easier to manage.
Results
FlexDeploy supported rapid growth for the organization and allowed them to deliver more consistently and quickly—and reliably—to their customer. With FlexDeploy, the Business Intelligence team was able to reduce or eliminate manual and scripted processes, which resulted in faster deployments.
Automation also led to standardization across environments, leading, in turn, to improved deployment accuracy, shrunk outage windows, and made deployments less person-dependent. It also reduced staff hours significantly, which were then allocated to more value-added project work that the Business Intelligence team loved – providing BI solutions – rather than doing manual or administrative work.
