oci_metering_computation_schedule
This resource provides the Schedule resource in Oracle Cloud Infrastructure Metering Computation service.
Returns the created schedule.
Example Usage
resource "oci_metering_computation_schedule" "test_schedule" {
#Required
compartment_id = var.compartment_id
name = var.schedule_name
result_location {
#Required
bucket = var.schedule_result_location_bucket
location_type = var.schedule_result_location_location_type
namespace = var.schedule_result_location_namespace
region = var.schedule_result_location_region
}
schedule_recurrences = var.schedule_schedule_recurrences
time_scheduled = var.schedule_time_scheduled
#Optional
defined_tags = {"foo-namespace.bar-key"= "value"}
description = var.schedule_description
freeform_tags = {"bar-key"= "value"}
output_file_format = var.schedule_output_file_format
query_properties {
#Required
date_range {
#Required
date_range_type = var.schedule_query_properties_date_range_date_range_type
#Optional
dynamic_date_range_type = var.schedule_query_properties_date_range_dynamic_date_range_type
time_usage_ended = var.schedule_query_properties_date_range_time_usage_ended
time_usage_started = var.schedule_query_properties_date_range_time_usage_started
}
granularity = var.schedule_query_properties_granularity
#Optional
compartment_depth = var.schedule_query_properties_compartment_depth
filter = var.schedule_query_properties_filter
group_by = var.schedule_query_properties_group_by
group_by_tag {
#Optional
key = var.schedule_query_properties_group_by_tag_key
namespace = var.schedule_query_properties_group_by_tag_namespace
value = var.schedule_query_properties_group_by_tag_value
}
is_aggregate_by_time = var.schedule_query_properties_is_aggregate_by_time
query_type = var.schedule_query_properties_query_type
}
saved_report_id = oci_data_safe_report.test_report.id
}
Argument Reference
The following arguments are supported:
compartment_id
- (Required) The customer tenancy.defined_tags
- (Optional) (Updatable) Defined tags for this resource. Each key is predefined and scoped to a namespace. See Resource Tags. Example:{"foo-namespace.bar-key": "value"}
description
- (Optional) (Updatable) The description of the schedule.freeform_tags
- (Optional) (Updatable) Simple key-value pair that is applied without any predefined name, type or scope. Exists for cross-compatibility only. See Resource Tags. Example:{"bar-key": "value"}
name
- (Required) The unique name of the user-created schedule.output_file_format
- (Optional) (Updatable) Specifies the supported output file format.query_properties
- (Optional) The query properties.compartment_depth
- (Optional) The depth level of the compartment.date_range
- (Required) Static or dynamic date rangedateRangeType
, which corresponds with type-specific characteristics.date_range_type
- (Required) Defines whether the schedule date range is STATIC or DYNAMIC.dynamic_date_range_type
- (Required when date_range_type=DYNAMIC)time_usage_ended
- (Required when date_range_type=STATIC) The usage end time.time_usage_started
- (Required when date_range_type=STATIC) The usage start time.
filter
- (Optional) The filter object for query usage.granularity
- (Required) The usage granularity. DAILY - Daily data aggregation. MONTHLY - Monthly data aggregation. Allowed values are: DAILY MONTHLYgroup_by
- (Optional) Aggregate the result by. For example: [ “tagNamespace”, “tagKey”, “tagValue”, “service”, “skuName”, “skuPartNumber”, “unit”, “compartmentName”, “compartmentPath”, “compartmentId”, “platform”, “region”, “logicalAd”, “resourceId”, “tenantId”, “tenantName” ]group_by_tag
- (Optional) GroupBy a specific tagKey. Provide the tagNamespace and tagKey in the tag object. Only supports one tag in the list. For example: [ { “namespace”: “oracle”, “key”: “createdBy” ]key
- (Optional) The tag key.namespace
- (Optional) The tag namespace.value
- (Optional) The tag value.
is_aggregate_by_time
- (Optional) Specifies whether aggregated by time. If isAggregateByTime is true, all usage or cost over the query time period will be added up.query_type
- (Optional) The query usage type. COST by default if it is missing. Usage - Query the usage data. Cost - Query the cost/billing data. Allowed values are: USAGE COST USAGE_AND_COST
result_location
- (Required) (Updatable) The location where usage or cost CSVs will be uploaded defined bylocationType
, which corresponds with type-specific characteristics.bucket
- (Required) (Updatable) The bucket name where usage or cost CSVs will be uploaded.location_type
- (Required) (Updatable) Defines the type of location where the usage or cost CSVs will be stored.namespace
- (Required) (Updatable) The namespace needed to determine the object storage bucket.region
- (Required) (Updatable) The destination Object Store Region specified by the customer.
saved_report_id
- (Optional) The saved report ID which can also be used to generate a query.schedule_recurrences
- (Required) Specifies the frequency according to when the schedule will be run, in the x-obmcs-recurring-time format described in RFC 5545 section 3.3.10. Supported values are : ONE_TIME, DAILY, WEEKLY and MONTHLY.time_scheduled
- (Required) The date and time of the first time job execution.
** IMPORTANT ** Any change to a property that does not support update will force the destruction and recreation of the resource with the new property values
Attributes Reference
The following attributes are exported:
compartment_id
- The customer tenancy.defined_tags
- Defined tags for this resource. Each key is predefined and scoped to a namespace. See Resource Tags. Example:{"foo-namespace.bar-key": "value"}
description
- The description of the schedule.freeform_tags
- Simple key-value pair that is applied without any predefined name, type or scope. Exists for cross-compatibility only. See Resource Tags. Example:{"bar-key": "value"}
id
- The OCID representing a unique shedule.name
- The unique name of the schedule created by the user.output_file_format
- Specifies the supported output file format.query_properties
- The query properties.compartment_depth
- The depth level of the compartment.date_range
- Static or dynamic date rangedateRangeType
, which corresponds with type-specific characteristics.date_range_type
- Defines whether the schedule date range is STATIC or DYNAMIC.dynamic_date_range_type
-time_usage_ended
- The usage end time.time_usage_started
- The usage start time.
filter
- The filter object for query usage.granularity
- The usage granularity. DAILY - Daily data aggregation. MONTHLY - Monthly data aggregation. Allowed values are: DAILY MONTHLYgroup_by
- Aggregate the result by. For example: [ “tagNamespace”, “tagKey”, “tagValue”, “service”, “skuName”, “skuPartNumber”, “unit”, “compartmentName”, “compartmentPath”, “compartmentId”, “platform”, “region”, “logicalAd”, “resourceId”, “tenantId”, “tenantName” ]group_by_tag
- GroupBy a specific tagKey. Provide the tagNamespace and tagKey in the tag object. Only supports one tag in the list. For example: [ { “namespace”: “oracle”, “key”: “createdBy” ]key
- The tag key.namespace
- The tag namespace.value
- The tag value.
is_aggregate_by_time
- Specifies whether aggregated by time. If isAggregateByTime is true, all usage or cost over the query time period will be added up.query_type
- The query usage type. COST by default if it is missing. Usage - Query the usage data. Cost - Query the cost/billing data. Allowed values are: USAGE COST USAGE_AND_COST
result_location
- The location where usage or cost CSVs will be uploaded defined bylocationType
, which corresponds with type-specific characteristics.bucket
- The bucket name where usage or cost CSVs will be uploaded.location_type
- Defines the type of location where the usage or cost CSVs will be stored.namespace
- The namespace needed to determine the object storage bucket.region
- The destination Object Store Region specified by the customer.
saved_report_id
- The saved report ID which can also be used to generate a query.schedule_recurrences
- Specifies the frequency according to when the schedule will be run, in the x-obmcs-recurring-time format described in RFC 5545 section 3.3.10. Supported values are : ONE_TIME, DAILY, WEEKLY and MONTHLY.state
- The schedule lifecycle state.system_tags
- Usage of system tag keys. These predefined keys are scoped to namespaces. See Resource Tags. Example:{"orcl-cloud.free-tier-retained": "true"}
time_created
- The date and time the schedule was created.time_next_run
- The date and time of the next job execution.time_scheduled
- The date and time of the first time job execution.
Timeouts
The timeouts
block allows you to specify timeouts for certain operations:
* create
- (Defaults to 20 minutes), when creating the Schedule
* update
- (Defaults to 20 minutes), when updating the Schedule
* delete
- (Defaults to 20 minutes), when destroying the Schedule
Import
Schedules can be imported using the id
, e.g.
$ terraform import oci_metering_computation_schedule.test_schedule "id"