📈Managing your instance
When operating your own Private Ellipsis Drive Instance, two system-level accounts are available:
Admin
Backoffice
The initial credentials for these accounts are provided in the configuration file during deployment.
These accounts provide access to the Ellipsis Drive management dashboard, which is available through the UI after login.
The dashboard consists of two main sections:
User management dashboard
System health dashboard
Kubernetes support
Ellipsis Drive runs in Docker containers and is compatible with Kubernetes.
Kubernetes-based scaling is supported but not required. Scaling can also be managed manually through additional virtual machines.
User dashboard
The user dashboard provides operational insights and controls, including:
Total storage usage per user
Quota management
Usage intensity and activity statistics
System health dashboard
The system health dashboard provides infrastructure monitoring, including:
Node status across the Ellipsis Drive cluster
CPU and memory usage per component
Storage capacity of raster and vector processing nodes
Admin vs backoffice accounts
The admin account has full access to all system content, regardless of user-defined permissions. This includes access to all datasets and resources within the instance.
The backoffice account is limited to system-level metadata, user management, and operational monitoring. It does not grant access to user data content.
Troubleshooting and scaling guidance
If your instance is not performing as expected, use the system health dashboard to identify the bottleneck and apply the appropriate scaling action.
High API CPU usage or slow API responses
Add additional virtual machines to the API pool
Raster storage is full or near capacity
Increase disk size or add an additional virtual machine to the raster storage pool.
Vector storage is full or near capacity
Increase disk size or add an additional virtual machine to the vector storage pool
Raster activation nodes are saturated
Add virtual machines to the raster activation pool
Vector activation nodes are saturated
Add virtual machines to the vector activation pool
Slow API responses without high CPU usage
Add a virtual machine to the cache storage pool
File storage is full or near capacity
Increase underlying storage volume
Last updated