Change Management Automation
Automating ServiceNow change management as part of our GitLab CI deployment workflow.
A Go platform that plugs GitLab CI into the ServiceNow API to automate enterprise change management. It cut release cycle time by 55%, removed manual change-task work, standardized deployments across teams, and was approved by our governance board for production.
Every deployment meant manual paperwork
Every deployment used to require an engineer to hand-create a ServiceNow change task, paste in the build details, update it mid-deploy, and close it afterward. It was repetitive, error-prone, inconsistent between teams, and slow.
The steps engineers repeated, every time
Make the change task disappear into the pipeline
I owned the architecture and implementation end to end, under one constraint: zero disruption. Teams keep shipping exactly as before. One manual push kicks it off, and everything after runs automatically inside the pipeline.
From a checklist to a single deploy
Two modules in one image, run as pipeline jobs
Two modules in one Docker image, run as GitLab CI jobs. Both validate their input and pull credentials before calling the ServiceNow API.
Both modules pull ServiceNow credentials from HashiVault at runtime, so secrets never live in the pipeline or the image. The image and the apps' build metadata come from Artifactory.
Catch bad input before it becomes a change task
The Open Task module validates the user and the change-task content before anything reaches ServiceNow, so bad input is caught in the pipeline instead of becoming a broken change task.
Choices that shaped the system
Built in Go
Go was the enterprise standard, which kept the tool maintainable and set it up for broader adoption across teams down the line.
Split into Open and Close modules
This mirrors how deployments behave. If a deploy fails, you don't want a pile of meaningless tasks, so close only runs on success.
Validate before creating tasks
Validating deployment input at the pipeline stage stops invalid change tasks from ever being created, removing a whole class of manual cleanup.
Runtime secrets via HashiVault
ServiceNow credentials are pulled from HashiVault at runtime, so they never live in the pipeline configuration or the Docker image.
A reusable image, dropped into existing pipelines
One Docker image dropped into a pipeline meant teams could adopt it with no onboarding and no change to how they deploy.
What it changed
What I took away
Design for adoption, not just functionality
The best automation is the kind users barely notice. Integrating seamlessly into existing deployment workflows made adoption straightforward across teams.
Validate early to reduce downstream risk
Catching invalid inputs before creating change tasks eliminated manual errors and improved the reliability of every deployment.
Build systems that standardize good practices
Automating repetitive processes did not just save time, it ensured deployments followed a consistent, auditable workflow every time.
Go
Docker
GitLab CI/CD
ServiceNow
HashiVault
Artifactory