Cloud VM Management Guide #

Quick reference for managing your UE5 build VM across different cloud providers.

Google Cloud Platform #

Create VM #

./scripts/create-gcp-vm.sh YOUR_PROJECT_ID us-central1-a

Stop VM (save costs) #

gcloud compute instances stop ue5-builder-spiderlily --zone=us-central1-a

Start VM #

gcloud compute instances start ue5-builder-spiderlily --zone=us-central1-a

Connect via RDP #

# Get external IP
gcloud compute instances describe ue5-builder-spiderlily --zone=us-central1-a --format="get(networkInterfaces[0].accessConfigs[0].natIP)"

# Reset password if needed
gcloud compute reset-windows-password ue5-builder-spiderlily --zone=us-central1-a --user=builduser

Delete VM #

gcloud compute instances delete ue5-builder-spiderlily --zone=us-central1-a

Check costs #

# View billing
gcloud billing accounts list
gcloud billing projects describe YOUR_PROJECT_ID

AWS EC2 #

Create VM #

# Via AWS Console or CLI
aws ec2 run-instances \
  --image-id ami-xxxxx \
  --instance-type c6i.4xlarge \
  --key-name your-key \
  --security-group-ids sg-xxxxx \
  --block-device-mappings DeviceName=/dev/sda1,Ebs={VolumeSize=500,VolumeType=gp3}

Stop Instance #

aws ec2 stop-instances --instance-ids i-xxxxx

Start Instance #

aws ec2 start-instances --instance-ids i-xxxxx

Get Windows Password #

aws ec2 get-password-data --instance-id i-xxxxx --priv-launch-key your-key.pem

Terminate Instance #

aws ec2 terminate-instances --instance-ids i-xxxxx

Azure #

Create VM #

az vm create \
  --resource-group ue5-builder-rg \
  --name ue5-builder \
  --image Win2022Datacenter \
  --size Standard_D16s_v5 \
  --admin-username builduser \
  --admin-password 'YourPassword123!'

Stop VM (deallocate) #

az vm deallocate --resource-group ue5-builder-rg --name ue5-builder

Start VM #

az vm start --resource-group ue5-builder-rg --name ue5-builder

Get IP #

az vm show -d --resource-group ue5-builder-rg --name ue5-builder --query publicIps -o tsv

Delete VM #

az vm delete --resource-group ue5-builder-rg --name ue5-builder --yes

Cost-Saving Tips #

Run Only During Work Hours #

GCP Scheduled Stop/Start:

# Stop at 6 PM (18:00)
gcloud compute instances add-metadata ue5-builder-spiderlily --zone=us-central1-a \
  --metadata=shutdown-script="shutdown /s /f /t 0"

# Use Cloud Scheduler to start/stop on schedule

AWS Scheduled Actions:

# Create schedules via AWS Console or EventBridge
aws autoscaling put-scheduled-action \
  --auto-scaling-group-name ue5-builder-asg \
  --scheduled-action-name stop-evening \
  --recurrence "0 18 * * *" \
  --min-size 0

Use Preemptible/Spot Instances #

GCP Preemptible (70-90% discount):

gcloud compute instances create ue5-builder-preemptible \
  --preemptible \
  --machine-type n2-standard-16 \
  ...

AWS Spot Instances (50-90% discount):

aws ec2 request-spot-instances \
  --instance-count 1 \
  --type persistent \
  --launch-specification file://spec.json

Caveat: Can be terminated at any time - not good for long builds!

Resize When Not Building #

Downsize to smaller instance during idle periods:

GCP:

gcloud compute instances set-machine-type ue5-builder-spiderlily \
  --zone=us-central1-a \
  --machine-type=n2-standard-4

AWS:

aws ec2 modify-instance-attribute \
  --instance-id i-xxxxx \
  --instance-type c6i.xlarge

Monitoring Costs #

GCP #

# View current month costs
gcloud billing accounts describe YOUR_BILLING_ACCOUNT

# Set budget alerts in Console
# https://console.cloud.google.com/billing/budgets

AWS #

# View costs
aws ce get-cost-and-usage \
  --time-period Start=2025-10-01,End=2025-10-31 \
  --granularity MONTHLY \
  --metrics BlendedCost

Azure #

# View costs
az consumption usage list --start-date 2025-10-01 --end-date 2025-10-31

Automation Scripts #

Auto-start on build trigger (optional) #

Create a GitHub Action that starts the VM before building:

- name: Start GCP VM
  run: |
    gcloud compute instances start ue5-builder-spiderlily --zone=us-central1-a
    sleep 60  # Wait for boot

Auto-stop after build #

Add to workflow:

- name: Stop VM
  if: always()
  run: |
    gcloud compute instances stop ue5-builder-spiderlily --zone=us-central1-a

Backup and Snapshots #

GCP Disk Snapshot #

gcloud compute disks snapshot ue5-builder-spiderlily \
  --snapshot-names=ue5-backup-$(date +%Y%m%d) \
  --zone=us-central1-a

AWS EBS Snapshot #

aws ec2 create-snapshot \
  --volume-id vol-xxxxx \
  --description "UE5 builder backup"

Create Machine Image (for quick restore) #

GCP:

gcloud compute images create ue5-builder-image \
  --source-disk=ue5-builder-spiderlily \
  --source-disk-zone=us-central1-a

Troubleshooting #

VM won't start #

• Check quotas (GCP: IAM & Admin > Quotas)
• Verify billing account is active
• Check instance logs in cloud console

Can't connect via RDP #

• Verify firewall rules allow port 3389
• Check external IP hasn't changed
• Reset Windows password

High costs #

• Check VM is stopping properly after builds
• Verify no orphaned disks/snapshots
• Use preemptible/spot instances for testing
• Consider smaller VM size

Best Practices #

1. Tag resources for easy tracking and cost allocation
2. Set up billing alerts to avoid surprises
3. Use snapshots before major changes
4. Stop VM when not actively building
5. Monitor usage weekly
6. Test with smaller VM first, then scale up if needed