Lakehouse Optimizer Incidents and Notifications Configuration
The Lakehouse Optimizer empowers users to monitor and improve their Lakehouse infrastructure by configuring incidents of interest that identify inefficiencies in cost, performance, and operational metrics. Incidents are displayed in the Incidents section of the app, providing actionable insights to optimize resource utilization and expenditures.
Incident Configuration Steps
Under the Settings menu, select Settings / Incident Policies
Incidents can be defined within Subscriptions, Workspaces, Workflows, All Purpose Compute, Delta Live Tables, SQL Warehouses, Pools, and Job Compute areas.
Each incident created has its own incident policy
Select the area of interest to create an incident under (ex. Workflow).
Select category - Cost Control or Performance
Select sub-category from dropdown menu (ex. Over Provisioning)
Select sub-option within Sub-Category (ex. Cluster CPU Over Provisioning)
Select +ADD button under Incident Rules section
Specify threshold for the incident rule (ex. Cluster CPU under 60% for any run) and save changes.
a. multiple rules can be set for each incident policy (ex. if a user wants an incident created each time Cluster CPU for any run is under 60%, 65%, or 70%)
b. When an incident rule has been met, an Incident Ticket is created automatically (and corresponding incident will appear in the Incidents view.
Every incident rule for a given Incident Policy automatically has email notifications turned on. However, by default no email group is tied to the incident rule, thus email notifications will not be sent until an email group(s) is selected for a given incident rule.
To assign an email group(s) to an incident rule, select the Email Group dropdown and select email group(s) (ex. Leadership Team, lho-support)
a. You will now see Leadership Team and lho-support displayed for Email Groups. Now save your changes to the incident rule.
To configure additional incident policies, you can navigate to Incident Policies under Settings or select the dropdown at the top of the screen and choose a different area of interest.
The following incidents are configurable in LHO:
Entity | Incident |
---|---|
Subscriptions | Monthly Cost above Threshold |
Workspaces | Monthly Cost above Threshold |
Workflows | Monthly Cost above Threshold |
Workflows |
|
Workflows |
|
Workflows |
|
Workflows | Bad Skew |
Workflows | Disk Spillage |
Workflows | Run Failure |
Workflows | Job with All Purpose Clusters |
Delta Live Tables | Over Provisioning
|
Delta Live Tables | Under Provisioning
|
Delta Live Tables | Update Failure |
Delta Live Tables | Monthly Cost above Threshold |
All Purpose Clusters | Monthly Cost above Threshold |
All Purpose Clusters | Auto Shutdown Timeout
|
All Purpose Clusters | Total Idle Time above Threshold |
All Purpose Clusters |
|
All Purpose Clusters |
|
Pools | Auto Shutdown Timeout
|
Email Group Configuration Steps
To create a new email group, select Settings / Email Notifications
Select +Add Group for Custom Groups
Specify group name (ex. Blueprint Test) and Save
Select +Add Email for the email group
Specify an email account and Add
Repeat process to include additional email accounts for a group
See steps #8 and #9 under Incident Configuration Steps above on how to assign an email group to an incident policy.