Skip to main content
For detailed cluster installation instructions, see Kubernetes Installation Guide. For troubleshooting, check the Troubleshooting Section.

Managing your Clusters with Qovery

Creating a Cluster

To create a new cluster:
  1. Open the Qovery Console
  2. Navigate to the Cluster page in the left menu
  3. Click Add Cluster
  4. Follow the provider-specific guide for your cloud platform:

Overview

Once your cluster is created, the cluster card displays comprehensive information about your cluster’s configuration and health: Cluster Details:
  • Deployed nodes and their current statuses
  • Instance type configuration (min/max instances)
  • Reserved resources: CPU, memory, and disk capacity
  • Current Kubernetes version with upgrade warnings if applicable
  • Last deployment timestamp and status
  • Karpenter nodepool grouping (if using Karpenter)
  • Per-node resource allocation
Cluster Overview

Statuses

Qovery tracks cluster health using two status categories:

Cluster Statuses

StatusDescription
RunningCluster is healthy and operational
WarningMinor issues detected, requires attention
ErrorCritical issues, intervention required
Status unavailableCluster unreachable or offline

Deployment Statuses

StatusDescription
Deployment QueuedDeployment is waiting in the queue
DeployingDeployment is currently in progress
Last Deployment FailedThe most recent deployment encountered errors
Last Deployment SucceededThe most recent deployment completed successfully

Performing Actions on Clusters

Available cluster operations and their cloud provider compatibility:
ActionAWSAzureGCPScalewayDescription
UpdateRedeploy cluster after configuration changes
StopTemporarily pause cluster (provider charges may still apply)
RestartResume a stopped cluster
DeletePermanently remove cluster
To perform an action:
  1. Navigate to your cluster page
  2. Click the action dropdown menu
  3. Select the desired action
  4. Track progress through cluster logs

Updating a Cluster

Redeploys your cluster after making configuration changes through the Qovery Console or API.

Stopping a Cluster

Temporarily pauses your cluster to reduce costs.
Some cloud provider charges may continue even when the cluster is stopped (e.g., storage, reserved IPs, load balancers).

Restarting a Cluster

Resumes a stopped cluster and redeploys all services.

Deleting a Cluster

Permanently removes your cluster. You have three deletion options:
  1. Default (Recommended): Complete cleanup
    • Deletes all Qovery-managed resources
    • Removes cloud provider infrastructure
    • No residual costs
  2. Delete on cloud provider and Qovery configuration: Manual database handling
    • Deletes cluster from cloud provider
    • Removes Qovery configuration
    • Requires manual database cleanup
  3. Delete Qovery configuration only: Keep cloud resources
    • Removes Qovery management only
    • Cloud infrastructure continues running
    • Continue managing manually in cloud console
Cluster deletion is permanent and irreversible. Ensure you have backups before proceeding.

Additional Actions

Audit Logs:
  • Access complete cluster activity history
  • Track configuration changes and deployments
  • Available for compliance and troubleshooting
Cluster ID:
  • Retrieve unique cluster identifier
  • Required for API operations and support tickets
Kubeconfig:
  • Download Kubernetes configuration file
  • Enable direct kubectl access
  • Use for advanced debugging and management
Security Best Practice: Kubeconfig provides cluster-admin access. Store securely and never commit to version control.

Logs

Access cluster logs for troubleshooting and monitoring:
  1. Navigate to your cluster page
  2. Click Logs or View Logs
  3. Use the tab system to view:
    • Cluster Info: General cluster information
    • Error Details: Specific error messages and stack traces
Access Cluster Logs
Successful Operations:
Successful Cluster Logs
Error Logs:
Cluster Error Logs

Custom Domain and Wildcard TLS (Beta)

By default, Qovery assigns a subdomain under qovery.io to your cluster services. However, Let’s Encrypt rate limits may affect certificate issuance. To use your own custom domain with wildcard TLS certificates, Qovery supports Cloudflare integration.

Requirements

  • A domain managed by Cloudflare
  • Cloudflare API token with specific permissions

Setup Steps

1

Create Cloudflare API Token

  1. Log into your Cloudflare account
  2. Go to My ProfileAPI Tokens
  3. Click Create Token
  4. Use the Edit zone DNS template or create a custom token
  5. Grant these permissions:
    • Zone → DNS → Edit
    • Zone → Zone → Read
  6. Set Zone Resources to include your domain
  7. Create and copy the token
2

Configure in Qovery

  1. Navigate to your cluster settings
  2. Find the Custom Domain section
  3. Enter your domain (e.g., example.com)
  4. Paste your Cloudflare API token
  5. Save configuration
3

Update DNS

Qovery will provide DNS records to add to your Cloudflare account
This feature is currently in Beta. Additional DNS providers will be supported in future releases.

AWS Account Cleanup

If cluster deletion fails or leaves resources behind, you can manually clean up using AWS Console tags.

Manual Cleanup Steps

1

Open AWS Console

2

Search for Qovery Resources

  1. Select All supported resource types
  2. Add tag filter: ClusterId or QoveryCluster
  3. Search for resources
3

Review and Delete

  1. Review the list of tagged resources
  2. Manually delete each resource through AWS Console
  3. Common resources to check:
    • EC2 instances
    • EKS clusters
    • VPCs and networking components
    • Load balancers
    • EBS volumes
    • Security groups
AWS Console Cleanup
Manual cleanup should only be performed if automatic Qovery deletion fails. Always verify resources before deletion to avoid data loss.

FAQ

A cluster is a group of computing resources (nodes) that work together to run your applications. Learn more in our Basic Concepts guide.
Clusters provide several key benefits:
  • Kubernetes orchestration: Automatic container management, scaling, and healing
  • Disaster recovery: Built-in redundancy and failover capabilities
  • Auto-scaling: Automatically adjust resources based on demand
  • Environment isolation: Separate development, staging, and production workloads
  • Multi-region deployment: Deploy applications closer to your users
Instance availability varies by cloud provider:AWS: 400+ instance types available
  • General purpose, compute optimized, memory optimized, storage optimized, GPU instances
  • Supports both x86_64 and ARM architectures (Graviton)
Scaleway: Multiple instance ranges
  • From cost-effective shared instances to dedicated high-performance options
  • x86_64 architecture
GCP: GKE auto-pilot
  • Fully managed node provisioning
  • Automatic optimization based on workload requirements
  • x86_64 architecture
Azure: Various VM series
  • Multiple performance tiers and specialized workloads
  • x86_64 architecture
Architecture compatibility: Ensure your container images support your chosen architecture (x86_64 or ARM).
Qovery follows a careful upgrade process:
  1. Testing period: Each new Kubernetes version undergoes 1 month of internal testing
  2. Rollout: After testing, updates roll out to customers over 3 weeks
  3. Manual upgrades: Option available for customers who want to upgrade immediately
Never upgrade Kubernetes manually through your cloud provider console. This can cause incompatibilities with Qovery components. Always use Qovery-managed upgrades.
You’ll receive notifications when new versions are available for your clusters.
Security is our main concern. When vulnerabilities surface:
  1. Rapid identification: Continuous monitoring of security databases
  2. Impact assessment: Evaluate severity and affected systems
  3. Mitigation solutions: Develop and test patches
  4. Customer communication: Transparent notifications with recommended actions
  5. Deployment assistance: Guided remediation process
Critical security patches may be applied automatically after customer notification.
A mirroring registry is a private container registry that caches images from public registries, providing:
  • Faster image pulls (cached locally)
  • Protection against rate limits
  • Improved reliability and availability
  • Better control over image versions
Learn more in our Image Mirroring documentation.
An unavailable status can occur due to:Common Causes:
  • Cluster manually stopped through cloud provider
  • Cloud provider outage or service disruption
  • Network connectivity issues
  • Accidental deletion of critical resources
  • Control plane issues
Troubleshooting Steps:
  1. Check cloud provider status page for outages
  2. Verify cluster hasn’t been manually stopped in cloud console
  3. Review Qovery cluster logs for error messages
  4. Check cloud credentials are still valid
  5. Contact Qovery support if issue persists
Always manage clusters through Qovery Console to avoid status inconsistencies. Manual changes in cloud provider consoles can cause unexpected behavior.

Next Steps