Cross-team infrastructure project
Reworking a patch and deployment process across teams
I mapped out which services depended on which, then reorganized our patching so it ran in dependency order with as much in parallel as the systems could handle.
Shorter maintenance windows and better SLAs for customers
- Process Design
- Orchestration
- Cross-team
The problem
Our patching and deployment process was made up of a lot of individual automated jobs. Each one shut a service down, patched it, and brought it back up. The jobs worked fine on their own, but they weren’t organized around how the services depended on each other, so the whole maintenance window took longer than it needed to.
The approach
This was a cross-organization effort. I worked with an architect in the company to map out the service dependencies, which turned out to be the hard part. With that map, we reorganized the whole process: the services everything else depends on go first, and from there we run as many jobs at the same time as the systems can handle.
The result
The full process got a lot faster. We were able to shrink the maintenance window, which meant we could also offer customers better SLAs.
The lesson carries over to smaller businesses too. Sometimes the biggest win isn’t automating a new step, it’s figuring out the right order for the steps you already have.