Scaling DevOps Tooling for a Global Bank
1
Majority of developer requests now fulfilled within 1 day
0
No major incidents or service outages
Sub-second
Average response time for core service pages
THE CHALLENGE
Laying the foundation for global DevOps enablement
The client had selected a strong DevOps toolset including Jira, Confluence, GitHub, Jenkins, Sonatype Nexus and Black Duck but lacked the service model to manage them effectively at scale.
Key challenges included:
– High volumes of daily toolchain requests from developers
– Maintaining performance, uptime and stability of critical tools
– Delivering a responsive, automated and user-friendly support experience
– Reducing the manual workload on the service team while planning for long-term capability transfer
To succeed, the client needed a scalable, developer-centric service that could operate globally with minimal friction and maximum efficiency.
THE SOLUTION
Self-service DevOps, designed for scale
Catapult’s Enable team partnered closely with the client’s service management function to design and embed a self-sustaining DevOps service model.
Highlights of the solution:
· Embedded delivery model, with Catapult working alongside the internal service team for hands-on capability transfer
· Developer-facing service desk built in Jira for streamlined self-service request creation and tracking
· Automated request handling, reducing response times and service team workload
· Knowledge base creation in Confluence, enabling developers to resolve common queries independently
· Continuous enhancement of service offerings, based on usage patterns and user feedback
The model focused on automation, self-service and reliability, empowering developers to get what they needed, fast, without burdening the support team
THE RESULTS
One-day turnarounds. Sub-second response time
Catapult helped transform DevOps tooling into a scalable, resilient internal service, enabling global developer productivity while reducing overhead and ensuring long-term sustainability.
– Most developer requests fulfilled within one day
– No major incidents or service outages across critical toolchain components
– Sub-second average response times for core service pages
– Service operated by a lean team of ~15, down from 100+, with no telephone support required
– High user satisfaction, with developers able to access, configure and troubleshoot tools independently
