Schema:

ACCOUNT_USAGE

WAREHOUSE_EVENTS_HISTORY View¶

This Account Usage view can be used to return the events that have been triggered for the single-cluster and multi-cluster warehouses in your account in the last 365 days (1 year).

Supported events include:

  • Creating, dropping, or altering a warehouse, including resizing the warehouse.

  • Resuming or suspending a warehouse.

  • Resuming, suspending, or resizing a cluster in a warehouse (single-cluster and multi-cluster warehouses).

  • Stopping or starting additional clusters in a warehouse (multi-cluster warehouses only).

Columns¶

Column Name

Data Type

Description

TIMESTAMP

TIMESTAMP_LTZ

The timestamp when the event is triggered.

WAREHOUSE_ID

NUMBER

The unique warehouse ID (assigned by Snowflake) that corresponds to the warehouse name in your account.

WAREHOUSE_NAME

TEXT

The name of the warehouse in your account.

CLUSTER_NUMBER

NUMBER

If an event was triggered for a specific cluster in a multi-cluster warehouse, the number of the cluster (starting with 1) for which the event was triggered; if the event was triggered for all clusters in the warehouse or is not applicable for a single-cluster warehouse, NULL is displayed.

EVENT_NAME

TEXT

Name of the event. For the list of possible values, see EVENT_NAME Descriptions (in this topic).

EVENT_REASON

TEXT

The cause of the event. For the list of possible values, see EVENT_REASON Descriptions (in this topic).

EVENT_STATE

TEXT

State of an event that takes time to complete or can be canceled: STARTED, COMPLETED, PARTIALLY_COMPLETED.

USER_NAME [1]

TEXT

User who initiated the event.

ROLE_NAME [1]

TEXT

Role that was active in the session at the time the event was initiated.

QUERY_ID [1]

TEXT

Internal/system-generated identifier for the SQL statement.

[1] These fields display data only for queries executed from April 1, 2021 onwards.

Usage Notes¶

  • Latency for the view may be up to 3 hours.

  • Each event can produce multiple rows in the view for the following reasons:

    • Events often trigger additional, related events.

    • Events may take time to complete or may be canceled; if this occurs, the EVENT_STATE column indicates whether the event was started, completed, or canceled.

EVENT_REASON Descriptions¶

WAREHOUSE_AUTORESUME:

A suspended warehouse was resumed automatically because AUTO_RESUME is enabled for the warehouse and a SQL statement was submitted to the warehouse.

WAREHOUSE_RESUME:

A suspended warehouse was resumed manually by a user.

WAREHOUSE_AUTOSUSPEND:

A running warehouse was suspended automatically because AUTO_SUSPEND is enabled for the warehouse and the defined period of inactivity for AUTO_SUSPEND has passed.

WAREHOUSE_SUSPEND:

A running warehouse was suspended manually by a user.

WAREHOUSE_RESIZE:

A warehouse was resized.

RESOURCE_MONITOR_SUSPEND:

A warehouse was suspended because the credit quota for the resource monitor for the warehouse was reached.

MULTICLUSTER_SPINUP:

A new/suspended cluster was provisioned in a multi-cluster warehouse; not applicable to single-cluster warehouses.

MULTICLUSTER_SPINDOWN:

A running cluster was shut down in a multi-cluster warehouse; not applicable to single-cluster warehouses.

EVENT_NAME Descriptions¶

Warehouse-related Events:

CREATE_WAREHOUSE:

Triggered by the creation of a new warehouse, which can occur when a user manually creates a warehouse or when an account is provisioned and the default warehouse is automatically created in the account.

  • Cluster number: None (N/A)

  • Event state: COMPLETED

  • Event reason: None (N/A)

  • Billing impact: No billing activity for this event if the cluster is created with INITIALLY_SUSPENDED = TRUE; otherwise, metering starts when all compute resources are provisioned for the warehouse or the warehouse starts processing statements (if the warehouse starts processing statements before the resources are fully provisioned).

DROP_WAREHOUSE:

Triggered when an existing warehouse is dropped; all running statements on the warehouse are stopped and the compute resources are released.

  • Cluster number: None (N/A)

  • Event state: COMPLETED

  • Event reason: None (N/A)

  • Billing impact: Metering on the compute resources for the warehouse stops after all running statements complete.

ALTER_WAREHOUSE:

Triggered when the properties of an existing warehouse are changed, including resizing the warehouse. If the warehouse is resized, additional RESIZE_WAREHOUSE events are triggered. This event can also trigger RESUME_WAREHOUSE or SUSPEND_WAREHOUSE events.

  • Cluster number: None (N/A)

  • Event state: COMPLETED

  • Event reason: None (N/A)

  • Billing impact: The billing impact depends on the event(s) that were triggered.

RESIZE_WAREHOUSE:

Triggered by changing the size of a warehouse, which increases or decreases the compute resources in each cluster in the warehouse. For a running warehouse, this event also triggers a RESIZE_CLUSTER event for each cluster in the warehouse.

  • Cluster number: None (applies to all clusters)

  • Event state: STARTED, COMPLETED, or PARTIALLY_COMPLETED

  • Event reason: WAREHOUSE_RESIZE

  • Billing impact: Resizing a running warehouse adds or removes compute resources in each cluster in the warehouse; the new added resources start metering when they are provisioned, while the removed resources stop metering after they finished processing any running statements. Resizing a suspended warehouse does not provision any new resources for the warehouse.

RESUME_WAREHOUSE:

Triggered when a suspended warehouse is resumed or a new warehouse is created with INITIALLY_SUSPENDED = FALSE. This event also triggers a RESUME_CLUSTER event for each cluster in the warehouse.

  • Cluster number: None (applies to all clusters)

  • Event state: STARTED, COMPLETED, or PARTIALLY_COMPLETED

  • Event reason: WAREHOUSE_AUTORESUME or WAREHOUSE_RESUME

  • Billing impact: The metering begins once all the compute resources are provisioned for the warehouse.

SUSPEND_WAREHOUSE:

Triggered when a running warehouse is suspended. This event also triggers a SUSPEND_CLUSTER event for each cluster in the warehouse.

  • Cluster number: None (applies to all clusters)

  • Event state: STARTED, COMPLETED, or PARTIALLY_COMPLETED

  • Event reason: WAREHOUSE_AUTOSUSPEND or WAREHOUSE_SUSPEND

  • Billing impact: Metering on the compute resources for the warehouse stops after all running statements complete.

Cluster-related Events:

RESUME_CLUSTER:

Triggered when a suspended cluster is resumed.

  • Cluster number: Number of the resumed cluster (always 1 for a single-cluster warehouse)

  • Event state: COMPLETED

  • Event reason:

    • WAREHOUSE_AUTORESUME or WAREHOUSE_RESUME (single-cluster warehouses)

    • MULTICLUSTER_SPINUP (multi-cluster warehouses)

  • Billing impact: Metering starts on the compute resources for the cluster once they are provisioned.

SUSPEND_CLUSTER:

Triggered when a running cluster is suspended.

  • Cluster number: Number of the suspended cluster (always 1 for a single-cluster warehouse)

  • Event state: COMPLETED

  • Event reason:

    • WAREHOUSE_AUTOSUSPEND or WAREHOUSE_SUSPEND (single-cluster warehouses)

    • MULTICLUSTER_SPINDOWN (multi-cluster warehouses)

    • RESOURCE_MONITOR_SUSPEND (all warehouses)

  • Billing impact: Metering stops on the compute resources for the cluster after all running statements finish.

RESIZE_CLUSTER:

Triggered when a cluster is resized, usually as a result of resizing a warehouse.

  • Cluster number: Number of the resized cluster (always 1 for a single-cluster warehouse)

  • Event state: COMPLETED

  • Event reason:

    • WAREHOUSE_AUTORESUME or WAREHOUSE_RESUME (single-cluster warehouses)

    • MULTICLUSTER_SPINDOWN or MULTICLUSTER_SPINUP (multiple-cluster warehouses)

    • WAREHOUSE_RESIZE (all warehouses)

  • Billing impact: Depends on whether compute resources are added or removed due to the resizing:

    • Compute resources that are added start metering after they are provisioned.

    • Compute resources that are removed stop metering after all running statements finish.

SPINUP_CLUSTER:

Triggered when a cluster is started (multi-cluster warehouse only); usually happens when the mininimum/maximum cluster size is increased.

  • Cluster number: Number of the cluster that was started

  • Event state: STARTED, COMPLETED, or PARTIALLY_COMPLETED

  • Event reason: MULTICLUSTER_SPINUP or WAREHOUSE_RESIZE

  • Billing impact: Metering start on the compute resources for the cluster once they are provisioned.

SPINDOWN_CLUSTER:

Triggered when a running cluster is shut down (multi-cluster warehouse only); usually happens when the mininimum/maximum cluster size is decreased.

  • Cluster number: Number of the cluster that was shut down

  • Event state: STARTED, COMPLETED, or PARTIALLY_COMPLETED

  • Event reason: MULTICLUSTER_SPINDOWN or WAREHOUSE_RESIZE

  • Billing impact: Metering stops on the compute resources for the cluster after all running statements finish.

Language: English