Depending on your project location, the Platform is using the cloud provider's spare capacity resources at a significant discount. This option will be marked as Spot instances on the visual interface.
The setting can be applied at the project and task level, with the latter overriding the former.
Data Cruncher will still use on-demand instances when running, even if the Spot Instance feature is turned on for that project.
When creating a new project, the default setting for Spot Instances is On.
Only project administrators can activate Spot Instances within that project. Spot Instances can be activated while creating a project, or subsequently within project settings following the procedure below:
- Go to your project dashboard.
- Click the Settings tab.
- Under Execution Settings switch Spot Instances to On.
Note that settings at task level override project-level settings.
- Create a draft task
- On the draft task page, switch Spot Instances to On underneath the draft task title.
An executed task set to use Spot Instances will be labelled Spot Instances: On on the task page underneath the task title.
On the task stats page you get a more detailed view of the executed task, with a job-wise breakdown of the instances used. Jobs that ran on Spot Instances are labelled spot next to the instance name.
use_interruptible_instances parameter provides equivalent control over the Spot Instance setting as part of the following API calls:
Updated about a year ago