Graphs on Hypermode is currently in developer preview. New features are shipping weekly.

​
Region availability

Graphs on Hypermode are available in the following regions:
RegionLocation
iad1Washington, D.C., USA
pdx1Portland, Oregon, USA
fra1Frankfurt, Germany
Additional regions are planned to be available in the near future:
  • Dublin, Ireland
  • Singapore

​
Scaling

Graph compute is measured in gigabytes (GB) of memory, with 1 vCPU allocated per 4 gigabytes of memory. On Hypermode, you’re charged for the compute allocated to your graph, billed to the second. That means when your graph is scaled to zero, you’re only billed for the underlying storage.

​
Scale to zero

By default, new graphs scale to zero after five minutes of inactivity. When a new query or mutation is received, the graph scales up and process that request. You can configure the scaling behavior for your graph by increasing the idle period. Set the idle period to 0 to turn off scale to zero.

​
Autoscaling

Autoscaling is a feature that automatically scales your graph based on the compute and memory load. When under sustained load, the graph scales up to handle the additional load. When the load subsides, the graph scales down automatically. You can configure the autoscaling behavior for your graph by setting the min and max values. The min value is the minimum number of compute units the graph has, and the max value is the maximum number of compute units the graph can scale to.

​
Storage

You are billed only for the storage used by your graph. The storage is automatically provisioned and expanded based on the growth of data within your graph.

​
Backups

Your graph on Hypermode is automatically backed up every four hours, with backups stored with zonal isolation. To restore your graph to a previous state, please reach out to support via the console.
User control of backup frequency and restore is coming soon.