- Added configurable provisioning timeout for
databricks_mws_workspaces
, so that local DNS cache issues would be more tolerated. - Added
databricks_current_user
to simplify applying the same Terraform configuration by different users in the shared workspace for testing purposes. - Added client-side rate limiting to release the pressure on backend APIs and prevent client blocking (#465)
- Member usernames, group names and instance profile names in
databricks_group
data source are now sorted and providing consistent behavior between runs (#449) - Fixed redundant multiple mounting clusters (#445)
- Added optional parameter azure_environment to provider config which defaults to public (#437).
- Added databricks_service_principal resource.
skip_validation
fromdatabricks_instance_profile
was removed and is always set totrue
.- Added propagation of terraform version to
User-Agent
header, along with type of resource used. databricks_notebook
&databricks_dbfs_file
got newsource
field to specify location of a local file.databricks_notebook
can havelanguage
field optional, as long assource
is set to a file with.py
,.scala
,.sql
, or.r
extension.databricks_me
data source was added to representuser_name
,home
&id
of the caller user (or service principal).- Added validation for secret scope name in
databricks_secret
,databricks_secret_scope
anddatabricks_secret_acl
. Non-compliant names may cause errors. - Added databricks_spark_version data source.
- Fixed support for single node clusters support by allowing
num_workers
to be0
. - Fixed bug in destruction of IP access lists (#426).
- All resource imports are now making call to corresponding Databricks API by default (#471).
Behavior changes
- Removed deprecated
library_jar
,library_egg
,library_whl
,library_pypi
,library_cran
, andlibrary_maven
fromdatabricks_cluster
anddatabricks_job
in favor of more API-transparent library configuration block. - Removed deprecated
notebook_path
andnotebook_base_parameters
fromdatabricks_job
in favor of notebook_task configuration block. - Removed deprecated
jar_uri
,jar_main_class_name
, andjar_parameters
fromdatabricks_job
in favor of spark_jar_task configuration block. - Removed deprecated
python_file
andpython_parameters
fromdatabricks_job
in favor of spark_python_task configuration block. - Removed deprecated
spark_submit_parameters
fromdatabricks_job
in favor of spark_submit_task configuration block. - Removed deprecated
databricks_scim_user
resource in favor of databricks_user. - Removed deprecated
databricks_scim_group
resource in favor of databricks_group. - Removed deprecated
databricks_default_user_roles
data source in favor of databricks_group data source. - Removed deprecated
basic_auth
andazure_auth
provider configuration blocks in favor of documented authentication methods. format
,overwrite
, andmkdirs
were removed fromdatabricks_notebook
. To follow expected behavior of Terraform, notebooks are always overwritten.skip_validation
fromdatabricks_instance_profile
was removed and is always set totrue
for subsequent requests.databricks_mws_workspace
gotverify_workspace_runnning
removed and now validates all every deployment. In case deployment failed, it removes workspace that failed and returns error message with explanation.default_tags
were removed fromdatabricks_instance_pool
.disk_spec
got new attributedisk_type
, that containsazure_disk_volume_type
andebs_volume_type
. This change is made to closer reflect API structure.databricks_notebook
&databricks_dbfs_file
gotcontent
attribute renamed tocontent_base64
and now share the same logic to work with local files.