Skip to content

Managing Deployments

This guide covers how to manage your deployed capabilities, including updates, deletions, and monitoring.

Viewing Deployed Capabilities

Installation History

Access your deployment history from the Welcome screen by clicking View Installation History:

Installation History

The Installation History displays: - Capability Name: The deployed capability - Version: Deployed version number - Environment: Target environment (dev/staging/prod) - Region: AWS region - Deployment Date: When the deployment occurred - Status: Current deployment status

Capability Selection Screen

From the main capability screen, you can see all your deployments:

Capability Selection

Installed Capabilities (green section) - Currently deployed capabilities - Active environments - Last deployment timestamp - Update and Delete actions

Available Capabilities (blue section) - Capabilities ready to install - Can install to new environments

Updating Deployments

When to Update

Update a deployment when: - New capability version is available - Configuration changes are needed - Infrastructure updates are required - Security patches need to be applied

Update Process

  1. Go to Capability Selection

Capability Selection

  1. Find your installed capability in the green section

  2. Click "Update"

  3. Review current configuration

  4. Environment and region are pre-populated

  5. Enter AWS credentials

AWS Configuration

  1. Verify permissions

Permission Verification

  1. Select deployment strategy

  2. Execute update

Deployment Progress

Update Strategies

Strategy Use When Impact
Update Existing Stacks are healthy Minimal downtime
Clean Deploy Stacks are in failed state Full recreation, data loss

Update vs. Clean Deploy

Update Existing: - Preserves existing resources - Updates only changed components - Minimal downtime - Preserves data

Clean Deploy: - Deletes all existing stacks - Creates fresh infrastructure - Causes downtime - May lose data (backup first!)

Deleting Deployments

When to Delete

Delete a deployment when: - Capability is no longer needed - Environment is being decommissioned - Starting fresh after issues - Cost optimization

Delete Process

  1. Go to Capability Selection

Capability Selection

  1. Find your installed capability

  2. Click "Delete"

  3. Confirm deletion (this is destructive!)

  4. Enter AWS credentials

AWS Configuration

  1. Verify permissions

  2. Execute deletion

What Gets Deleted

The deletion process removes:

Frontend Resources: - S3 buckets (and contents) - CloudFront distributions - WAF rules - ACM certificates - Route53 records

Backend Resources: - ECS services and clusters - Load balancers - VPC and networking - DynamoDB tables - S3 data buckets - Secrets Manager secrets - CloudWatch logs - IAM roles

Data Backup

Before deleting, backup: - DynamoDB tables - S3 bucket contents - Secrets Manager values - CloudWatch logs

Backup Commands

# Backup DynamoDB table
aws dynamodb create-backup \
  --table-name capability-prod-data \
  --backup-name capability-backup-$(date +%Y%m%d)

# Sync S3 bucket
aws s3 sync s3://capability-prod-data s3://capability-backup-data

# Export secrets
aws secretsmanager get-secret-value \
  --secret-id capability-prod-secrets \
  --query SecretString \
  --output text > secrets-backup.json

Monitoring Deployments

Deployment Health

After deployment, verify service health:

Post Verification

Health Status Indicators: - ✅ Healthy - Service is operational - ⚠️ Degraded - Service has issues - ❌ Unhealthy - Service is down

CloudWatch Metrics

Monitor your deployment health via CloudWatch:

ECS Metrics: - CPU utilization - Memory utilization - Running task count - Service health

ALB Metrics: - Request count - Response time - Error rates (4xx, 5xx) - Active connections

CloudFront Metrics: - Requests - Bytes downloaded - Error rate - Cache hit ratio

Setting Up Alarms

Create CloudWatch alarms for:

Metric Threshold Action
ECS CPU > 80% Scale up
ECS Memory > 80% Scale up
ALB 5xx errors > 1% Alert
ALB latency > 2s Alert

Viewing Logs

ECS Container Logs: 

aws logs tail /ecs/capability-prod --follow

Application Logs: 

aws logs tail /aws/application/capability-prod --follow

Deployment States

Understanding Stack States

State Description Actions Available
CREATE_COMPLETE Successfully created Update, Delete
UPDATE_COMPLETE Successfully updated Update, Delete
CREATE_IN_PROGRESS Being created Wait
UPDATE_IN_PROGRESS Being updated Wait
DELETE_IN_PROGRESS Being deleted Wait
ROLLBACK_COMPLETE Creation failed Delete only
UPDATE_ROLLBACK_COMPLETE Update failed Delete or Clean Deploy
DELETE_FAILED Deletion failed Manual cleanup

Handling Failed States

ROLLBACK_COMPLETE: 1. Review CloudFormation events for error 2. Fix the underlying issue 3. Delete the stack 4. Redeploy

UPDATE_ROLLBACK_COMPLETE: 1. Review what caused the update to fail 2. Choose Clean Deploy strategy 3. Or manually fix and retry update

DELETE_FAILED: 1. Check which resources failed to delete 2. Manually delete stuck resources 3. Retry stack deletion

Multi-Environment Management

Environment Strategy

Environment Purpose Resources
dev Development Minimal, cost-optimized
staging Pre-production Production-like
prod Production Full resources, HA

Deploying to Multiple Environments

Use the Environment Configuration screen to select your target:

Environment Configuration

Recommended Workflow:

  1. Deploy to dev first
  2. Test new features

  3. Validate configuration

  4. Promote to staging

  5. Integration testing

  6. Performance testing

  7. Deploy to production

  8. After staging validation
  9. During maintenance window

Environment Isolation

Each environment has: - Separate AWS resources - Unique domain names - Independent scaling - Isolated data

Cost Management

Cost Breakdown

Component Dev Staging Prod
CloudFront $5 $10 $50+
ALB $20 $20 $20
ECS Fargate $10 $50 $200+
NAT Gateway $0 $45 $45
S3 $1 $5 $20+
DynamoDB $5 $10 $50+
Total ~$41 ~$140 ~$385+

Cost Optimization Tips

Development: - Use FARGATE_SPOT for ECS - Disable NAT Gateway (use public subnets) - Set auto-scaling minimum to 0 - Use on-demand DynamoDB billing

Production: - Use Reserved Capacity for predictable workloads - Enable S3 lifecycle policies - Optimize CloudFront caching - Right-size ECS tasks

Cleaning Up Unused Resources

Regularly review and delete: - Unused ECS task definitions - Old CloudWatch log groups - Orphaned S3 buckets - Unused security groups

Troubleshooting Deployments

Common Issues

Deployment Stuck: 1. Check CloudFormation events 2. Look for resource creation failures 3. Check service limits

Health Checks Failing:

Post Verification

  1. Verify security group rules
  2. Check container logs
  3. Validate health check endpoint

DNS Not Resolving: 1. Wait for propagation (up to 48 hours) 2. Verify Route53 records 3. Check hosted zone configuration

SSL Certificate Issues: 1. Verify domain ownership 2. Check DNS validation records 3. Wait for certificate validation

Getting Help

  1. Check deployment logs in the installer

Deployment Logs

  1. Review CloudFormation events in AWS Console
  2. Check CloudWatch logs for application errors
  3. Contact support with deployment ID and logs

Managing Kubernetes (EKS) Deployments

If you deployed using the EKS (Kubernetes) option, your capability runs as a NexusAICapability custom resource managed by the Kubernetes Operator.

Viewing Kubernetes Deployments

# List all deployed capabilities
kubectl get nexuscapabilities -A

# Short form
kubectl get tc -A

# Get detailed status
kubectl describe tc <capability-name> -n <namespace>

Checking Application Pods

# List pods in capability namespace
kubectl get pods -n <capability>-<environment>

# Example: nexus-ai-dev
kubectl get pods -n nexus-ai-dev

# View pod logs
kubectl logs -f deployment/<capability>-frontend -n <namespace>
kubectl logs -f deployment/<capability>-backend -n <namespace>

Updating Kubernetes Deployments

# Scale replicas
kubectl patch tc <name> -n <namespace> --type=merge \
  -p '{"spec":{"backend":{"replicas":3}}}'

# Update image version
kubectl patch tc <name> -n <namespace> --type=merge \
  -p '{"spec":{"backend":{"image":"new-image:tag"}}}'

# Or edit directly
kubectl edit tc <name> -n <namespace>

Deleting Kubernetes Deployments

# Delete capability (operator handles cleanup)
kubectl delete tc <name> -n <namespace>

Note: AWS resources (DynamoDB, S3) are deleted or retained based on the deletionPolicy in the capability spec.

Operator Management

# Check operator status
kubectl get pods -n nexus-system

# View operator logs
kubectl -n nexus-system logs -f deployment/nexus-operator

# Restart operator
kubectl -n nexus-system rollout restart deployment/nexus-operator

Kubernetes-Specific Monitoring

Application Metrics: 

# CPU and memory usage
kubectl top pods -n <namespace>

# Events
kubectl get events -n <namespace> --sort-by='.lastTimestamp'

Service URLs: 

# Get LoadBalancer URLs
kubectl get svc -n <namespace>

# Frontend URL
kubectl get svc <capability>-frontend -n <namespace> \
  -o jsonpath='{.status.loadBalancer.ingress[0].hostname}'

For detailed Kubernetes deployment instructions, see Kubernetes Deployment Guide.

Best Practices

Deployment

  • ✅ Always test in dev/staging first
  • ✅ Use consistent naming conventions
  • ✅ Document configuration changes
  • ✅ Backup data before updates

Monitoring

  • ✅ Set up CloudWatch alarms
  • ✅ Enable detailed monitoring
  • ✅ Review logs regularly
  • ✅ Track costs monthly

Security

  • ✅ Rotate credentials regularly
  • ✅ Review IAM permissions
  • ✅ Enable AWS Config rules
  • ✅ Use AWS Security Hub

Kubernetes-Specific

  • ✅ Monitor operator health regularly
  • ✅ Use deletionPolicy: Retain for production data
  • ✅ Configure resource limits in capability spec
  • ✅ Set up Pod Disruption Budgets for HA

Next Steps