This repository also provides organization-wide reusable workflows under
.github/workflows/:
python-container.yml— install, test and build a Python container;node-container.yml— install, typecheck/build and build a Node container;java-container.yml— run Maven tests and build a Java container;helm-validate.yml— lint and render a Helm chart, then validate the rendered manifest is non-empty without requiring a cluster connection.container-cleanup.yml— inspect and retain protected/recent GHCR versions.
Repositories call a template from a thin workflow in their own repository:
jobs:
verify:
uses: devsecopslonghn/.github/.github/workflows/python-container.yml@master
with:
image_name: example-service
publish_image: false
secrets: inheritGHCR publication uses the built-in github.token; no external registry
credential is required. A caller can enable it with publish_image: true,
registry: ghcr.io and a full image name such as
ghcr.io/devsecopslonghn/example-service. The template publishes both the
commit SHA and latest; pull requests still build without publishing. Pin
callers to a release tag or commit SHA when the template contract is stable.
Application workflows publish immutable image tags and update the target Helm repository. Argo CD is the only component that reconciles Kubernetes; GitHub Actions does not apply manifests directly.
GHCR packages must be either public for anonymous cluster pulls or private with
an explicitly provisioned Kubernetes imagePullSecret backed by a credential
that has package read access. Registry credentials are never committed to a
repository or embedded in workflow logs.
Container cleanup is a separate scheduled workflow. It protects latest, the
image tags referenced by the target Helm values, and keeps the 20 most recent
remaining versions. Manual dispatch is dry-run by default; scheduled runs are
the only automatic deletion path. The workflow uses the repository
GITHUB_TOKEN, which must have package administration access for the package.
I build CI/CD platforms, deployment automation, and reliability tooling for banking and enterprise systems.
Currently focused on Core Banking, Open Banking, Payment Systems, OpenShift, GitHub Actions, GitLab CI, Ansible, Helm, and database release automation.
Website | LinkedIn | GitHub | Email
DevOps Engineer based in Ho Chi Minh City, Vietnam, with 3+ years of experience in regulated financial environments.
My work is centered around secure and auditable delivery:
- CI/CD pipeline engineering for multi-environment deployments
- OpenShift and Kubernetes deployment automation
- Reusable GitHub Actions, GitLab CI, and release governance
- Helm-based application delivery for modular upgrades
- Database CI/CD with versioned migrations and rollback planning
- Security scanning integration with SonarQube, Aqua Scanner, Coverity, and BlackDuck
- Monitoring and incident visibility using Grafana, Prometheus, Datadog, Zabbix, and Kibana
- Managing CI/CD patterns across large banking application portfolios
- Automating deployment workflows for Core Banking, Open Banking, and Payment systems
- Improving release safety through approval gates, audit logs, and rollback strategies
- Building platform workflows that connect Git, Ansible, ITSM, and Microsoft Teams
- Contributed to Temenos T24 R25 go-live delivery in Vietnam through pipeline automation and deployment standardization
- Helped reduce a major banking upgrade window from weeks to days through repeatable CI/CD and Helm-based delivery
- Redesigned JBoss deployment topology for an AML platform to remove avoidable deployment downtime
- Built and maintained automation patterns for WebSphere, JBoss, OpenShift, WSO2 API Manager, and database deployments
- Supported production-grade reliability practices with monitoring, alerting, cleanup automation, and incident response
| Area | Tools |
|---|---|
| CI/CD | GitLab CI, GitHub Actions, and reusable workflow automation |
| Containers | Docker, Kubernetes, OpenShift OCP, Helm |
| Automation | Ansible, Terraform, Bash, Python |
| Database Delivery | Flyway, SQL versioning, migration and rollback scripts |
| Monitoring | Grafana, Prometheus, Datadog, Zabbix, Kibana |
| Security | SonarQube, Aqua Scanner, Coverity, BlackDuck |
| Artifact Management | Nexus, JFrog Artifactory |
| Cloud | AWS EC2, S3, RDS, ECS, Auto Scaling, Spot Instances |
Personal DevOps portfolio built with Astro and Tailwind CSS. It presents my experience, CV, production delivery background, and DevOps project notes.
Personal automation client for VPN connection workflows.
Simple database design exercise for categories, questions, answer choices, users, and scoring.
Public-safe summaries of real delivery patterns I work with: branch-based release strategy, approval gates, OpenShift rollout design, database migration governance, and monitoring-driven operations.
For DevOps, CI/CD, platform automation, or banking delivery discussions:
- Email: longhn0710@gmail.com
- LinkedIn: linkedin.com/in/longhn0710