v2.32.0
版本发布时间: 2023-08-17 22:12:36
hasura/graphql-engine最新发布版本:v2.44.0(2024-10-08 18:36:54)
Changelog
Highlights
Support for SQL Server stored procedures is now GA
We are delighted to announce that our Support for SQL Server stored procedures is now GA. The Hasura GraphQL Engine can now track read-only SQL Server Stored Procedures using the Logical Models feature.
This allows you to define custom GraphQL types directly in the Hasura GraphQL Engine, to define the result set returned from the stored procedure, and automatically generate a GraphQL root field for querying it. See the related docs.
Behaviour changes
- A new metadata inconsistency error will be raised when inherited roles are configured where the parent roles have different column select permissions for the same table/logical model when the data source does not support the column redaction feature.
- Postgres/Citus/Cockroach/MSSQL/BigQuery support column redaction.
- Athena/Snowflake/MySQL/MariaDB/Oracle/MongoDB do not currently support column redaction.
Bug fixes and improvements
Server
- Adds configuration option to specify a custom header for JWT authentication.
- Fixes an issue where an MS SQL Server source was unable to track a relationship on a foreign key when the referenced table resided in a different schema.
- Improves handling of inherited roles and fixes a bug for where ordering and filtering would cause the query to fail in MS SQL Server.
- Simplifies Postgres aggregation predicate SQL generation by removing unnecessary
AND
andOR
s inside SQLWHERE
clauses. - Allow Stored Procedures to be tracked and untracked using the
bulk_atomic
metadata API command. (Cloud / Enterprise edition only) - Allow
where
filters to refer to nested fields in non-relational databases that store data as structured documents such as MongoDB. (Cloud / Enterprise edition only) - Adds a new Prometheus metric
hasura_websocket_message_write_time
and fixes existinghasura_websocket_message_queue_time
to exclude websocket message write time. (Cloud / Enterprise edition only) - Allows filter permissions to refer to fields of nested objects in document databases such as MongoDB. (Cloud / Enterprise edition only)
- Support
order_by
for fields of nested objects for document databases ingraphql-engine
with a compatible data connector agent such as MongoDB. This requires a version of a compatible agent that includes support forColumnSelector
. (Cloud / Enterprise edition only) - Respect individually-disabled OpenTelemetry export telemetry types when an endpoint URL is set. (Enterprise edition only)
Console
- Improves the selection of array data types during column creation in the Console.
- Improves suggested relationships with more natural and intuitive naming.
- Improves function tracking on the Console for Postgres and SQL Server. This can be enabled via
Settings
>Feature Flags
>Improved Tracking UI for Postgres and SQL Server
. - Improves Remote Schema relationships UX by adding a search option for types, better organizing configuration options, and easing navigation when mapping fields.
- Fixes an issue in the Console that prevented new relationships from being created if there were inconsistent sources in the Metadata.
- Editing of Logical Models is now supported via the Console.
- Display only local relationships in the permissions builder as permissions using remote relationships are not yet supported.
- Fixes issue while creating RESTified endpoints in the following scenarios: when the
graphql-default
naming convention is enabled, custom table name GraphQL customization is set for a table, or Remote Schemas are present. - Fixes remote database relationship visualization in the Console
- Adds minor UX improvements to the Stored Procedure UI i.e. better error handling, empty states, prompts.
- Fixes issue where columns were not displaying for materialized views in the new Relationships UI.
- Fixes JWT token analyzer in API explorer.
- Fixes GraphiQL loading indefinitely in case no GraphQL schema is returned.
Data Connectors
-
hasura/graphql-data-connector
docker image will run using non-root user (UID: 1001) and group (GID: 1001) by default. (Enterprise edition only)
1、 cli-hasura-darwin-amd64 75.41MB
2、 cli-hasura-darwin-arm64 75.43MB
3、 cli-hasura-linux-amd64 23.5MB
4、 cli-hasura-linux-arm64 22.05MB
5、 cli-hasura-windows-amd64.exe 69.27MB