Deploying a Service

Automatic Deployment (CI/CD)

Most services deploy automatically when code is pushed to GitHub. The CI pipeline:

1. Builds a Docker image
2. Pushes to harbor.ayinza.dev
3. Runs docker service update --image on the Swarm

No manual intervention needed if the pipeline is set up for the service.

Manual Deployment

When you need to deploy manually (e.g., hotfix, pipeline down, or infrastructure service update):

1. SSH into Azure (the Swarm Leader):

   ssh kaks@4.180.181.86

2. Pull the new image (if needed):

   docker pull harbor.ayinza.dev/library/service-name:tag

3. Update the service:

   docker service update \
     --image harbor.ayinza.dev/library/service-name:tag \
     --with-registry-auth \
     dev_service-name

4. Verify the deployment:

   docker service ps dev_service-name
   docker service logs -f dev_service-name --tail 50

Redeploying an Entire Stack

If you need to redeploy all services in a stack (e.g., after changing the stack YAML):

DEV

docker stack deploy -c /home/kaks/stacks/infrastructure.yml dev
docker stack deploy -c /home/kaks/stacks/services.yml dev

UAT

docker stack deploy -c /home/kaks/stacks/uat-infrastructure.yml uat
docker stack deploy -c /home/kaks/stacks/uat-services.yml uat

Shared

docker stack deploy -c /home/kaks/stacks/shared-infrastructure.yml shared

Nginx

docker stack deploy -c /home/kaks/stacks/nginx.yml dev

Note

docker stack deploy is idempotent — it only updates services that have changed. It won’t restart services that haven’t changed.

Force Restarting a Service

Sometimes a service needs a restart without an image change:

docker service update --force dev_admin-service

This triggers a rolling restart of all replicas.

Checking Deployment Status

# See current state of all replicas
docker service ps dev_service-name

# Watch logs in real-time
docker service logs -f dev_service-name --tail 100

# See all services and their status
docker service ls