Destroying Deployments

When you’re done with a deployment, use the /destroy command to properly clean up cloud resources and local configuration.

Destroying a deployment is permanent and irreversible. All data on the server will be lost.

What Gets Destroyed

The destroy process removes: Cloud Resources:

VPS server instance
Associated SSH keys from cloud provider
Public IP address
All data on the server

Local Files:

Deployment configuration (~/.clawcontrol/deployments/NAME/config.json)
Deployment state (~/.clawcontrol/deployments/NAME/state.json)
SSH key pair (~/.clawcontrol/deployments/NAME/ssh/)
Entire deployment directory

What Doesn’t Get Destroyed:

Your cloud provider account
API keys stored in ClawControl
Other deployments
Templates

Destroying a Deployment

Launch destroy command

From ClawControl’s main menu:

/destroy

ClawControl lists all deployments:

/destroy - Select a deployment to destroy

WARNING: This action cannot be undone!

> my-openclaw-server     [deployed]
  old-test-server        [failed]
  dev-server             [initialized]

Arrow keys to select | Enter to destroy | Esc to go back

Select deployment

Use arrow keys to select the deployment you want to destroy and press Enter.

You can destroy deployments in any state: initialized, deployed, configuring, or failed.

Confirm destruction

ClawControl shows what will be deleted and asks for confirmation:

Confirm Destruction

You are about to destroy:
Deployment: my-openclaw-server
Server IP: 159.89.123.45

This will permanently delete:
• The VPS server (if deployed)
• All data on the server
• Local configuration files
• SSH keys

Type the deployment name to confirm:

Confirm: _

Type the exact deployment name to confirm:

Confirm: my-openclaw-server

Press Enter to proceed.

Destruction in progress

ClawControl connects to your cloud provider and removes resources:

Destroying deployment...
Deleting server and cleaning up resources...

This typically takes 10-30 seconds.

Destruction complete

When finished:

Deployment Destroyed

The deployment "my-openclaw-server" has been permanently deleted.

Press any key to return to home

The deployment no longer appears in /list or /status.

What Happens During Destruction

Delete cloud server

ClawControl calls your cloud provider’s API to delete the VPS:Hetzner:

DELETE /servers/{server_id}

DigitalOcean:

DELETE /droplets/{droplet_id}

The server is immediately shut down and scheduled for deletion.

Remove SSH key

The SSH key uploaded during deployment is removed from your cloud provider:Hetzner:

DELETE /ssh_keys/{key_id}

DigitalOcean:

DELETE /account/keys/{key_id}

Clean local files

All local files are deleted:

rm -rf ~/.clawcontrol/deployments/my-openclaw-server/

This includes:

Configuration and state files
SSH key pair
Deployment directory

Special Cases

Destroying Initialized Deployments

For deployments that were never deployed (status: initialized):

No cloud resources to delete
Only local files are removed
Use /list → [D]elete instead (faster)

/list → Select deployment → Press D

Destroying Failed Deployments

For deployments that failed during provisioning:

ClawControl attempts to delete any created resources
Even if the server doesn’t exist, cleanup succeeds
Local files are always removed

If a deployment failed early, the server may not exist. ClawControl handles this gracefully and completes the cleanup.

Destroying Multiple Deployments

To destroy multiple deployments:

Run /destroy for each one individually
Confirm each destruction

There’s no bulk destroy option for safety.

Troubleshooting Destruction

API Key Expired or Invalid

Problem: Can’t delete server because API key is invalid Solution:

Manually delete the server from your cloud provider’s dashboard:
- Hetzner: Cloud Console → Servers → Delete
- DigitalOcean: Droplets → Your Server → Destroy
Run /destroy again - it will skip cloud deletion and clean up local files

Server Already Deleted

Problem: Server was manually deleted from cloud provider Solution: Run /destroy normally. ClawControl detects the server is gone and completes local cleanup.

Local Files Deleted Manually

Problem: Deployment directory was deleted manually Solution:

Server still exists in cloud (costing money)
Manually delete from cloud provider’s dashboard
Delete SSH key from cloud provider (search for “clawcontrol-*”)

Always use /destroy instead of manually deleting files. Manual deletion leaves cloud resources running and costing money.

Destruction Hangs

Problem: Destruction process is stuck Solutions:

Wait 2-3 minutes (cloud providers can be slow)
Press Ctrl+C to cancel
Check cloud provider dashboard to see if server was deleted
Try /destroy again

Cost Implications

Hourly Billing

Most cloud providers bill by the hour: Hetzner:

Hourly billing, rounded up
CPX11: ~$0.007/hour
Destroy before hour ends to save money

DigitalOcean:

Hourly billing with monthly cap
s-1vcpu-2gb: $0.018/hour ($ 12/month cap)
Destroy anytime to stop charges

When to Destroy

Destroy immediately if:

Testing is complete
Deployment failed and you’re starting over
Switching to different configuration
No longer needed

Keep running if:

Production bot in active use
Long-running experiments
Shared with team
Cost is acceptable

For development and testing, destroy deployments when not in use to save costs. Recreate when needed (takes ~5-10 minutes).

Before You Destroy

Backup Data

If you need to preserve any data:

Connect via SSH

/ssh

Select your deployment.

Identify important files

# OpenClaw configuration
~/.openclaw/config.json
~/.openclaw/.env

# Logs
journalctl -u openclaw > openclaw-logs.txt

# Custom data
/path/to/your/data

Download files

From your local machine:

scp -i ~/.clawcontrol/deployments/NAME/ssh/id_ed25519 \
  root@SERVER_IP:~/.openclaw/config.json \
  ./backup-config.json

Export Configuration

To recreate the deployment later:

Configuration is in ~/.clawcontrol/deployments/NAME/config.json

Copy before destroying:

cp ~/.clawcontrol/deployments/my-openclaw-server/config.json ~/backups/

Or fork the deployment to preserve settings:

/list → Select deployment → Press F

Notify Users

If others use this bot:

Notify in Telegram/Slack that bot is shutting down
Provide timeline for new instance (if applicable)
Share new bot credentials when redeployed

After Destruction

Verify Cloud Resources Deleted

Check your cloud provider dashboard: Hetzner:

Go to Cloud Console
Navigate to Servers
Verify server is gone
Check SSH Keys section

DigitalOcean:

Go to Droplets
Verify droplet is gone
Check Settings → Security → SSH Keys

Check Billing

Monitor your cloud provider billing:

Hetzner: Check Cloud Console → Billing
DigitalOcean: Check Billing → Current Usage

Verify charges stopped after destruction.

Clean Up Tailscale (Optional)

If you used Tailscale:

Go to Tailscale Admin
Find the destroyed server
Click the ”…” menu → Delete

Alternative: Delete vs Destroy

ClawControl has two deletion commands:

/destroy

When to use: Deployed servers
Deletes:
  - Cloud VPS server
  - Cloud SSH keys
  - Local configuration
  - Local SSH keys
Safety: Type deployment name to confirm
Time: 10-30 seconds

/list → [D]elete

When to use: Initialized deployments only
Deletes:
  - Local configuration
  - Local SSH keys
Safety: Type deployment name to confirm
Time: Instant

Use /destroy for any deployment that was deployed (even if failed). Use /list → [D]elete only for deployments that were never deployed.

Best Practices

Development Workflow

Create deployment: /new
Deploy: /deploy
Test functionality
Destroy when done: /destroy
Repeat with different configuration

Cost for 2 hours of testing:

Hetzner CPX11: ~$0.02
DigitalOcean basic: ~$0.04

Production Workflow

Create deployment: /new
Deploy: /deploy
Test thoroughly
Monitor with /status and /logs
Keep running
Destroy only when replacing or retiring

Multi-Instance Strategy

dev-server:     Create/destroy frequently
staging-server: Keep running during development
prod-server:    Keep running continuously

Only destroy production when:

Migrating to new infrastructure
No longer needed
Major configuration change required

Recovery

Once a deployment is destroyed, it CANNOT be recovered. The server and all its data are permanently deleted.

If you destroyed by mistake:

Server data: Lost permanently
Configuration: Gone if not backed up
Recreation: Deploy fresh instance with similar config

To minimize risk:

Always confirm the deployment name carefully
Keep templates of common configurations
Document API keys externally
Back up important data regularly

Next Steps

First Deployment - Create a new deployment
Using Templates - Quickly recreate destroyed deployments
Managing Deployments - Fork before destroying to preserve config

Get Started

Core Concepts

Guides

Cloud Providers

Configuration

Destroying Deployments

What Gets Destroyed

Destroying a Deployment

What Happens During Destruction

Special Cases

Destroying Initialized Deployments

Destroying Failed Deployments

Destroying Multiple Deployments

Troubleshooting Destruction

API Key Expired or Invalid

Server Already Deleted

Local Files Deleted Manually

Destruction Hangs

Cost Implications

Hourly Billing

When to Destroy

Before You Destroy

Backup Data

Export Configuration

Notify Users

After Destruction

Verify Cloud Resources Deleted

Check Billing

Clean Up Tailscale (Optional)

Alternative: Delete vs Destroy

/destroy

/list → [D]elete

Best Practices

Development Workflow

Production Workflow

Multi-Instance Strategy

Recovery

Next Steps

Build docs developers (and LLMs) love

Get Started

Core Concepts

Guides

Cloud Providers

Configuration

​What Gets Destroyed

​Destroying a Deployment

​What Happens During Destruction

​Special Cases

​Destroying Initialized Deployments

​Destroying Failed Deployments

​Destroying Multiple Deployments

​Troubleshooting Destruction

​API Key Expired or Invalid

​Server Already Deleted

​Local Files Deleted Manually

​Destruction Hangs

​Cost Implications

​Hourly Billing

​When to Destroy

​Before You Destroy

​Backup Data

​Export Configuration

​Notify Users

​After Destruction

​Verify Cloud Resources Deleted

​Check Billing

​Clean Up Tailscale (Optional)

​Alternative: Delete vs Destroy

​/destroy

​/list → [D]elete

​Best Practices

​Development Workflow

​Production Workflow

​Multi-Instance Strategy

​Recovery

​Next Steps

Build docs developers (and LLMs) love

What Gets Destroyed

Destroying a Deployment

What Happens During Destruction

Special Cases

Destroying Initialized Deployments

Destroying Failed Deployments

Destroying Multiple Deployments

Troubleshooting Destruction

API Key Expired or Invalid

Server Already Deleted

Local Files Deleted Manually

Destruction Hangs

Cost Implications

Hourly Billing

When to Destroy

Before You Destroy

Backup Data

Export Configuration

Notify Users

After Destruction

Verify Cloud Resources Deleted

Check Billing

Clean Up Tailscale (Optional)

Alternative: Delete vs Destroy

/destroy

/list → [D]elete

Best Practices

Development Workflow

Production Workflow

Multi-Instance Strategy

Recovery

Next Steps