Data Source: oci_dataflow_application
This data source provides details about a specific Application resource in Oracle Cloud Infrastructure Data Flow service.
Retrieves an application using an applicationId
.
Example Usage
data "oci_dataflow_application" "test_application" {
#Required
application_id = oci_dataflow_application.test_application.id
}
Argument Reference
The following arguments are supported:
application_id
- (Required) The unique ID for an application.
Attributes Reference
The following attributes are exported:
application_log_config
- Logging details of Application logs for Data Flow Run.log_group_id
- The log group id for where log objects will be for Data Flow Runs.log_id
- The log id of the log object the Application Logs of Data Flow Run will be shipped to.
archive_uri
- A comma separated list of one or more archive files as Oracle Cloud Infrastructure URIs. For example,oci://path/to/a.zip,oci://path/to/b.zip
. An Oracle Cloud Infrastructure URI of an archive.zip file containing custom dependencies that may be used to support the execution of a Python, Java, or Scala application. See https://docs.cloud.oracle.com/iaas/Content/API/SDKDocs/hdfsconnector.htm#uriformat.arguments
- The arguments passed to the running application as command line arguments. An argument is either a plain text or a placeholder. Placeholders are replaced using values from the parameters map. Each placeholder specified must be represented in the parameters map else the request (POST or PUT) will fail with a HTTP 400 status code. Placeholders are specified asService Api Spec
, wherename
is the name of the parameter. Example:[ "--input", "${input_file}", "--name", "John Doe" ]
If “input_file” has a value of “mydata.xml”, then the value above will be translated to--input mydata.xml --name "John Doe"
class_name
- The class for the application.compartment_id
- The OCID of a compartment.configuration
- The Spark configuration passed to the running process. See https://spark.apache.org/docs/latest/configuration.html#available-properties. Example: { “spark.app.name” : “My App Name”, “spark.shuffle.io.maxRetries” : “4” } Note: Not all Spark properties are permitted to be set. Attempting to set a property that is not allowed to be overwritten will cause a 400 status to be returned.defined_tags
- Defined tags for this resource. Each key is predefined and scoped to a namespace. For more information, see Resource Tags. Example:{"Operations.CostCenter": "42"}
description
- A user-friendly description.display_name
- A user-friendly name. This name is not necessarily unique.driver_shape
- The VM shape for the driver. Sets the driver cores and memory.driver_shape_config
- This is used to configure the shape of the driver or executor if a flexible shape is used.memory_in_gbs
- The amount of memory used for the driver or executors.ocpus
- The total number of OCPUs used for the driver or executors. See here for details.
execute
- The input used for spark-submit command. For more details see https://spark.apache.org/docs/latest/submitting-applications.html#launching-applications-with-spark-submit. Supported options include--class
,--file
,--jars
,--conf
,--py-files
, and main application file with arguments. Example:--jars oci://path/to/a.jar,oci://path/to/b.jar --files oci://path/to/a.json,oci://path/to/b.csv --py-files oci://path/to/a.py,oci://path/to/b.py --conf spark.sql.crossJoin.enabled=true --class org.apache.spark.examples.SparkPi oci://path/to/main.jar 10
Note: If execute is specified together with applicationId, className, configuration, fileUri, language, arguments, parameters during application create/update, or run create/submit, Data Flow service will use derived information from execute input only.executor_shape
- The VM shape for the executors. Sets the executor cores and memory.executor_shape_config
- This is used to configure the shape of the driver or executor if a flexible shape is used.memory_in_gbs
- The amount of memory used for the driver or executors.ocpus
- The total number of OCPUs used for the driver or executors. See here for details.
file_uri
- An Oracle Cloud Infrastructure URI of the file containing the application to execute. See https://docs.cloud.oracle.com/iaas/Content/API/SDKDocs/hdfsconnector.htm#uriformat.freeform_tags
- Free-form tags for this resource. Each tag is a simple key-value pair with no predefined name, type, or namespace. For more information, see Resource Tags. Example:{"Department": "Finance"}
id
- The application ID.idle_timeout_in_minutes
- The timeout value in minutes used to manage Runs. A Run would be stopped after inactivity for this amount of time period. Note: This parameter is currently only applicable for Runs of typeSESSION
. Default value is 2880 minutes (2 days)language
- The Spark language.logs_bucket_uri
- An Oracle Cloud Infrastructure URI of the bucket where the Spark job logs are to be uploaded. See https://docs.cloud.oracle.com/iaas/Content/API/SDKDocs/hdfsconnector.htm#uriformat.max_duration_in_minutes
- The maximum duration in minutes for which an Application should run. Data Flow Run would be terminated once it reaches this duration from the time it transitions toIN_PROGRESS
state.metastore_id
- The OCID of Oracle Cloud Infrastructure Hive Metastore.num_executors
- The number of executor VMs requested.owner_principal_id
- The OCID of the user who created the resource.owner_user_name
- The username of the user who created the resource. If the username of the owner does not exist,null
will be returned and the caller should refer to the ownerPrincipalId value instead.parameters
- An array of name/value pairs used to fill placeholders found in properties likeApplication.arguments
. The name must be a string of one or more word characters (a-z, A-Z, 0-9, _). The value can be a string of 0 or more characters of any kind. Example: [ { name: “iterations”, value: “10”}, { name: “input_file”, value: “mydata.xml” }, { name: “variable_x”, value: “${x}”} ]name
- The name of the parameter. It must be a string of one or more word characters (a-z, A-Z, 0-9, _). Examples: “iterations”, “input_file”value
- The value of the parameter. It must be a string of 0 or more characters of any kind. Examples: “” (empty string), “10”, “mydata.xml”, “${x}”
pool_id
- The OCID of a pool. Unique Id to indentify a dataflow pool resource.private_endpoint_id
- The OCID of a private endpoint.spark_version
- The Spark version utilized to run the application.state
- The current state of this application.time_created
- The date and time the resource was created, expressed in RFC 3339 timestamp format. Example:2018-04-03T21:10:29.600Z
time_updated
- The date and time the resource was updated, expressed in RFC 3339 timestamp format. Example:2018-04-03T21:10:29.600Z
type
- The Spark application processing type.warehouse_bucket_uri
- An Oracle Cloud Infrastructure URI of the bucket to be used as default warehouse directory for BATCH SQL runs. See https://docs.cloud.oracle.com/iaas/Content/API/SDKDocs/hdfsconnector.htm#uriformat.