Manage Deployment
DIGIT uses Helm as its primary packaging and deployment tool, with all environment configuration stored declaratively in the DIGIT-DevOps Git repository. This GitOps approach means every deployment is version-controlled, reproducible, and auditable. Services are deployed via either Jenkins automation pipelines or direct Helm/Helmfile commands.
Deployment Approach
Packaging
Helm charts. Every DIGIT service is packaged as a Helm chart. Charts consist of templates (YAML manifests) and a values.yaml file containing configurable defaults.
Environment config
Environment-specific values are stored in deploy-as-code/helm/environments/<env>.yaml inside DIGIT-DevOps. Secrets are encrypted with SOPS before committing.
Orchestration
Helmfile. Manages multiple Helm releases declaratively. Supports diff, apply, sync, and destroy commands. Uses helm-diff and helm-secrets plugins.
CI/CD
Jenkins (self-hosted) or managed DevOps (AWS CodeDeploy, Azure DevOps, NIC DevOps). Jenkins pipelines trigger deployments from the DIGIT-DevOps repo.
Source of truth
Git (DIGIT-DevOps repository). No live environment changes without a corresponding Git commit. Drift is prevented by re-applying the Git state.
Helmfile Deployment Workflow
Clone the DIGIT-DevOps repository and switch to the LTS branch before running any deploy commands:
git clone https://github.com/egovernments/DIGIT-DevOps.git
cd DIGIT-DevOps
git checkout DIGIT-2.9LTS(Helmfile)
cd deploy-as-code/charts/monitoringInstall the required Helm plugins before the first deployment:
# Package manager for Kubernetes
curl https://raw.githubusercontent.com/helm/helm/main/scripts/get-helm-3 | bash
# Declarative Helm release manager
wget https://github.com/helmfile/helmfile/releases/latest/download/helmfile_linux_amd64 \
-O /usr/local/bin/helmfile && chmod +x /usr/local/bin/helmfile
# Shows diff between current and proposed state
helm plugin install https://github.com/databus23/helm-diff
# Decrypts SOPS-encrypted secrets automatically during deploy
helm plugin install https://github.com/jkroepke/helm-secrets
# Verify plugins
helm plugin listStandard deployment sequence for any environment:
1. Preview
helmfile -e <env> -f helmfile.yaml template
Renders Kubernetes manifests without applying. Review before committing.
2. Diff
helmfile -e <env> -f helmfile.yaml diff
Shows exactly what will change in the cluster vs current state.
3. Apply
helmfile -e <env> -f helmfile.yaml apply
Applies changes. Prompts for confirmation when using --interactive flag.
4. Sync
helmfile -e <env> -f helmfile.yaml sync
Forces full sync. Use when diff reports drift between Git state and cluster.
5. Destroy
helmfile --interactive destroy
Removes all releases defined in the helmfile. Requests confirmation before deleting.
Jenkins CI/CD Pipeline
Jenkins provides a GUI-driven deployment pipeline as an alternative to direct Helmfile commands. It is suited for teams that require approval gates, audit logs, and role-based deployment access.
Jenkins Master VM
4 vCore, 8 GB RAM. Manages job scheduling, user authentication, and build history.
Jenkins Slave VM
4 vCore, 8 GB RAM. Executes deployment jobs. Communicates with the Kubernetes cluster via kubeconfig.
Deployment jobs
Environment-specific jobs (e.g. deploy-to-dev, deploy-to-uat, deploy-to-prod) defined in seed-deployment-jobs in jenkins/values.yaml.
Build trigger
Paste a Docker image ID into the build job and click Build. Jenkins runs the Helm deployer against the target environment.
Service deployment
go run main.go deploy -e <env> -c '<service-list>' — Golang deployer script used to deploy one or more named services to a specific environment.
To deploy specific services using the Go deployer (alternative to full Helmfile sync):
Rollback & Version Management
Helm maintains a revision history for every release. If a deployment introduces a regression, roll back to the previous stable revision without redeploying from scratch.
Pre & Post Deployment Checklist
Pre-deploy
Confirm environment config (env.yaml) is committed and SOPS secrets are encrypted
Pre-deploy
Run helmfile diff and review all changes — no surprises in apply
Pre-deploy
Verify all pods in the target namespace are Running or Completed (no pending/crash-loop pods)
Pre-deploy
Confirm database backup has been taken before any schema-changing deployment
Pre-deploy
Notify stakeholders of planned maintenance window for Prod deployments
Post-deploy
Verify citizen portal is available at <hostname>/digit-ui/citizen
Post-deploy
Verify employee portal is available at <hostname>/digit-ui/employee
Post-deploy
Check pod status: kubectl get pods -n egov — all should be Running
Post-deploy
Run a smoke test API call against the API gateway to confirm routing
Post-deploy
Check Grafana dashboards for any spike in error rates or latency following deployment
Post-deploy
Confirm Alertmanager has not fired any critical alerts within 10 minutes of deployment
GitOps & Environment Configuration
All environment configuration lives in DIGIT-DevOps as code. Configuration changes follow the same Git workflow as application changes: branch, commit, review, merge, deploy.
environments/<env>.yaml
Environment-specific service values: db-host, ingress domain, service toggles, replica counts, resource limits.
environments/<env>-secrets.yaml
Encrypted secrets: DB passwords, Alertmanager credentials, Git private key, payment gateway keys. Encrypted with SOPS before commit.
charts/monitoring/monitoring-helmfile.yaml
Declares all monitoring stack releases (Prometheus, Grafana, Loki, Alertmanager) and their value overrides.
charts/backbone-services/
Kafka, ZooKeeper, PostgreSQL, ElasticSearch, Redis Helm charts and their values.
To update Loki to use S3 storage on AWS, modify the environment YAML and reapply:
Last updated
Was this helpful?