Client Overview
Our client is a fintech investment company based in Saudi Arabia. Their backend applications were hosted on on-premises IIS servers, managed manually with no CI/CD automation. As the user base grew, so did the infrastructure and deployment complexity—leading to operational inefficiencies and frequent disruptions.
Problem Statement
The client was deploying their .NET-based microservices manually by asking Techieonix engineers to rebuild and redeploy every time a code update was merged. This process was time-consuming, error-prone, and involved:
- Identifying which services were changed
- Manually building Docker images
- Pushing images to the OCI Container Registry
- Executing kubectl rollout to trigger updates
This created unnecessary delays and increased the client’s billable hours. Late-hour deployment requests from the client’s developers also made it hard to balance workload priorities across our DevOps team.
Requirement Gathering
The primary goal was to automate the deployment process for backend microservices, reduce manual involvement, and improve deployment speed and consistency. The solution needed to:
- Automatically detect which microservices were updated
- Build and push Docker images selectively
- Restart only the relevant Kubernetes deployments
- Maintain visibility and verification of successful rollouts
- Operate within OCI (Oracle Cloud Infrastructure) using their QA environment
Our Approach
We implemented a GitHub Actions-based CI/CD pipeline that fully automates the detection and deployment of changed services.
Git-based triggers
A new GitHub workflow watches for changes in the src/Backend/** directory of the cicd/serviceQA branch.
Service-to-deployment mapping
Each backend service is mapped to a corresponding Kubernetes deployment and image name.
Selective builds
Only modified services are built and deployed, minimizing resource consumption.
Kubernetes integration
kubectl rollout restart is executed programmatically for the affected services.
Deployment validation
Rollout status is monitored for each service to confirm successful deployment.
Implementation (Deployment Pipeline Steps)
Change Detection
GitHub Actions detect changes using git diff logic based on code pushed to a specific QA branch.
Build Phase
Only the updated services are built into Docker images.
Push Phase
Docker images are pushed to OCI Container Registry using retry logic to ensure reliability.
Deploy Phase
Corresponding Kubernetes deployments are restarted using kubectl rollout restart.
Verification
The rollout status of each deployment is monitored to ensure successful and stable deployment.
| Microservice | Kubernetes Deployment | Docker Image |
|---|---|---|
| Cis.Auth.Web | auth-service | corp-auth-service-qa:latest |
| Cis.Dashboard.Web | dashboard-service | corp-dashboard-service-qa:latest |
| ...and 20+ others | ||
Technologies Used
Oracle Cloud Infrastructure (OCI)
Kubernetes (OKE)
GitHub Actions
Docker
OCI Container Registry
WSL
Visual Studio CodeOutcome
Deployments are now fully automated post-merge, cutting build times by deploying only changed services. This reduced DevOps costs, eliminated late-night bottlenecks, and enabled faster, safer feature releases. Each rollout is monitored for stability, ensuring a reliable QA environment.
