Valohai provides automatic tracking of underutilized resources to help you optimize the efficiency of your machine learning efforts.
If the peak utilization is low, you may be able to reduce the resources allocated to your executions, money and time by using cheaper, more readily available instances.
Valohai automatically highlights executions that had less than 50% peak utilization of any resource. This is currently a fixed threshold, but you can specify which resources you are interested under Project > Notifications settings discussed further below.
Valohai also only reports on executions that took more than a minute to finish.
Note
If you want another threshold, please contact your Valohai representative for us to gather requirements for upcoming features.
Underutilization Indicators
Alerts
Alerts lightweight messages that are displayed on the execution page.
They can’t be turned off but can be hidden on execution lists.
Alerts as shown on the execution details page.
You may turn on showing the alerts on the execution listings.
- Column settings; Alerts column visibility can be changed here
- Alerts column; click to navigate to the full list of alerts
Notifications
Notifications can be received via e.g. in-app messages, email, or Slack.
There are two types of notifications:
- Personal notifications are sent to the user who configured the notification.
- Project notifications are sent to the target channel e.g. Slack or webhook.
You can configure both of these notifications in the project settings.
- Project settings page
- Notifications tab
- Personal notifications view
- Turn on/off in-app messages of that type
- Turn on/off email notifications of that type
- Change scope of the notification; do we report about your or all executions
Contrary to personal notifications, project notifications are shared with all users.
Note
Setting up project notifications requires configuring a channel as the receiver. This can be done under the Channels tab in the project notifications settings shown above.