Automation HealthPRODUCTION
_AutomationInstance
Automation run-level health: status, duration, schedule context and first error detail.
Retention: Past 6 monthsScope: Single business unitTime: UTC
Trust & provenancePRODUCTION
Evidence tierPRODUCTIONOfficial sources1Other sources0Last verifiedNot recorded
Evidence tier describes my hands-on relationship with the content. Official sources describe platform contracts/documentation. Production observations should not automatically be treated as official limits.
Starter query
Start from something valid.
SELECT
au.AutomationCustomerKey,
au.AutomationInstanceID,
au.AutomationInstanceStatus,
au.AutomationInstanceStartTime_UTC,
au.AutomationInstanceEndTime_UTC
FROM _AutomationInstance auRelationships
Where it connects.
Schema
Fields
| Field | Type | Flags | Notes |
|---|---|---|---|
MemberID | Number | Business Unit identifier. | |
AutomationName | Text(400) | Automation name. | |
AutomationCustomerKey | Text(400) | INDEXED | Unique automation customer key. |
AutomationType | Text(9) | NULLABLE | Starting source: Schedule, File Drop or Trigger. |
AutomationStepCount | Number | Number of automation steps. | |
AutomationInstanceID | Text / UUID | INDEXED | Unique identifier for this automation run. |
FilenameFromTrigger | Text(4000) | NULLABLE | File that started a file-drop/triggered automation. |
AutomationInstanceScheduledTime_UTC | Date | NULLABLE | Scheduled start time for scheduled automations. |
AutomationInstanceStartTime_UTC | Date | NULLABLE | Actual run start time. |
AutomationInstanceEndTime_UTC | Date | NULLABLE | Run end time; null while running or when skipped. |
AutomationInstanceStatus | Text(400) | Run status such as QueuedFile, Initialized, Executing, Stopped, Complete or Error. | |
AutomationInstanceActivityErrorDetails | Text(4000) | NULLABLE | First activity error detail for the run. |
Operational notes
Available in real time.
Salesforce notes this view can generate one row per automation notification recipient; deduplicate by AutomationInstanceID when building health views.
Use duration trends and Skipped status as health signals, not only Error.
Verified sources