hasura/graphql-engine v2.27.0-beta.1

on GitHub

Changelog

Highlights

Support aggregations on computed fields

Postgres and Postgres-compatible backends now support comparison and numeric aggregation functions (such as min, max, sum, stddev and so on) on computed fields. (close #3238)

Bug fixes and improvements

Server

• Add postgres_get_table_info, citus_get_table_info, and cockroach_get_table_info Metadata APIs.
• Array literals can now be used in permission checks for Data Connector-based data sources. Array literals are typically used when using the _in and _nin operators.
• Prefixed *_get_table_info metadata APIs will now return a pair for scalar types when handling Data Connector-backed sources: the first being the type according to the backend, and the second being the corresponding GraphQL type.
• Add support for root-field user-defined functions to the Data Connector protocols.
• Expose and visualize metrics using hasura_cache_request_count to gain insights into your cache's performance. Learn more. (Enterprise edition only)
• Add trigger_name and source_name labels to existing Event Trigger metrics. (Cloud / Enterprise edition only)
• Add a new hasura_source_health Prometheus metric, corresponding to /healthz/sources. (Cloud / Enterprise edition only)
• Enhance Logical Models to allow nested arrays and arrays of scalars to be expressed. Logical Model scalar types should still parse as before. (Cloud / Enterprise edition only)
• Validation of Postgres native queries will no longer reject column names containing uppercase letters. (Cloud / Enterprise edition only)

Console

• Fix cron trigger invocation logs status
• Show the correct array relationship details in the new relationship UI.
• Increase the accuracy of the request time and size in the GraphiQL network information bar.
• Allow setting tolerance time for cron triggers from the Console.
• Prevent negative values in database connection fields on Console.
• Improve the performance of bulk table tracking for all databases.

Data Connectors

• Fix issue where required JDBC parameters where not being added to mutation requests. In particular, this affected MySQL mutations where allowMultiQueries was not present in the supplied JDBC string. (Cloud / Enterprise edition only)
• Add a new query-generating strategy for Snowflake. The new strategy is optimized internally by the Snowflake engine. It also restores the full compliment of Hasura permissions, which were limited under the previous strategy. (Cloud / Enterprise edition only)
• Fix errors in Snowflake queries caused by tables with non-uppercase names. (Cloud / Enterprise edition only)