Releases: hasura/graphql-engine
v2.39.1
Changelog
This is a patch release for v2.39
Bug fixes and improvements
Build
- Fix bug where connections to read replicas would persist even after they were removed from the data source config. (Cloud / Enterprise edition only)
v2.39.0
Changelog
Bug fixes and improvements
Server
- The server now logs timing information on error when possible, e.g. when an SQL query fails.
- Generate an info message instead of a warning when the GC is forced to run.
- Add
operation_name
andparameterized_query_hash
labels to thehasura_graphql_requests_total
Prometheus metric. (Cloud / Enterprise edition only) - Support env variable template to set the OpenTelemetry status (fix #10177) (Cloud / Enterprise edition only)
- Fix bug causing unstable connections for projects with multiple sources using the same SSL certificates. (Enterprise edition only)
Console
- Fix values not being suggested in dropdowns of the permissions builder if only 1 option exists
v2.39.0-beta.1
Changelog
Bug fixes and improvements
Server
- The server now logs timing information on error when possible, e.g. when an SQL query fails.
- Generate an info message instead of a warning when the GC is forced to run.
- Add
operation_name
andparameterized_query_hash
labels to thehasura_graphql_requests_total
Prometheus metric. (Cloud / Enterprise edition only) - Support env variable template to set the OpenTelemetry status (fix #10177) (Cloud / Enterprise edition only)
Console
- Fix values not being suggested in dropdowns of the permissions builder if only 1 option exists
v2.38.1
Changelog
This is a patch release for v2.38
Bug fixes and improvements
Build
- Update the Red Hat UBI base images to receive latest security updates
v2.38.0
Changelog
Behaviour changes
-
Add automatic quoting to Native Query parameters for Snowflake. (Cloud / Enterprise edition only)
Note that this may affect existing Native Queries in contexts where query authors have manually added quotes around parameter injection sites; removing the manual quotes should resolve this issue.
Bug fixes and improvements
Server
-
Add the ability to force refresh of dynamic db connection strings. (Cloud / Enterprise edition only)
When the
HASURA_SECRETS_BLOCKING_FORCE_REFRESH_URL
env var is defined, then on retry Hasura will post this payload to the URL:{"filename": <path>}
Crucially it is assumed that the Hasura server only returns after the referenced secret file has been updated.
-
Ignore dynamic headers
x-real-ip
,x-forwarded-for
,true-client-ip
while generating cache keys. (Cloud / Enterprise edition only)
Console
- Fix console crash when using Native Queries in an array relationship.
- Add ClickHouse and Trino database logos. Fix an issue where some logos may not have been displaying correctly for data-connector backed databases.
Data Connectors
- In the JVM connection pool, validate connections when borrowing them. This prevents the issue where, for example, the DB server is terminated and the pool is unaware that the connection is in an invalid state, leading to errors.
- Inline query parameters in output generated by the GraphiQL
Analyze
button (i.e. the Explain API). Also, use dialect-specific SQL based on the current connector rather than a default/generic dialect.
v2.38.0-beta.1
Changelog
Bug fixes and improvements
Server
- Ignore dynamic headers
x-real-ip
,x-forwarded-for
,true-client-ip
while generating cache keys. (Cloud / Enterprise edition only)
Console
- Fix console crash when using Native Queries in an array relationship.
- Add ClickHouse and Trino database logos. Fix an issue where some logos may not have been displaying correctly for data-connector backed databases.
Data Connectors
- Inline query parameters in output generated by the GraphiQL
Analyze
button (i.e. the Explain API). Also, use dialect-specific SQL based on the current connector rather than a default/generic dialect.
v2.37.1
Changelog
This is a patch release for v2.37
Bug fixes and improvements
Server
- Fix output of
error
field for asynchronous actions with 2xx response status.
Data connectors
- Remove explicit NULL ordering when generating ORDER BY statements.
Some dialects such as MySQL do not supportORDER BY ... NULLS FIRST/LAST
The translation toORDER BY CASE WHEN
causes the optimizer to be unable to use indices.
v2.37.0
Changelog
Bug fixes and improvements
Server
- When adding a data connector agent, the agent URI can now be sourced from an environment variable (fix #9777).
- Allow ignoring the
errors
field if bothdata
anderrors
are present in the Remote Schema response. Can be configured via the serverHASURA_GRAPHQL_REMOTE_SCHEMA_PRIORITIZE_DATA
env var or theremote-schema-prioritize-data
flag. - Remove additional
internal
error response field being returned in Async Actions when dev mode is false. - Ensure Logical Models do not create empty GraphQL object types.
- Add a new environment variable
HASURA_GRAPHQL_CONFIGURED_HEADER_PRECEDENCE
to specify header precedence when delivering payload to webhook in actions and input validations. - BigQuery filters now support
ilike
andnilike
for case-insensitive versions oflike
andnlike
.
Console
- Also prettify the query variables on hitting the GraphiQL Prettify button (fix #9968).
- Sort the columns in the filter and sort column selectors alphabetically in the data browser (fix #9966).
- Show table schema names in the new relationships UI for PostgreSQL and MS SQL Server.
Data Connectors
- Fix multi-schema issue for simple queries in Snowflake. (Cloud / Enterprise edition only)
- Fix aggregation limit for Snowflake, now aggregate limits are treated separately from row-based limits which can come through from permission-based limits. (Cloud / Enterprise edition only)
- Add support for the following additional JDBC auth mechanisms: (Cloud / Enterprise edition only)
- AWS
- password managed by the AWS Secrets Manager
- IAM based access for AWS (available for MySQL, MariaDB and PostgreSQL)
- Snowflake
- encrypted private keys
- AWS
Build
- Update the Ubuntu and Red Hat UBI base images to receive latest security updates
v2.36.3
Changelog
This is a patch release for v2.36
Bug fixes and improvements
Server
- Fix query caching when database connection templates are involved. Session variables and HTTP headers used in the templates are now included in the cache key generation. (Cloud / Enterprise edition only)
v2.37.0-beta.1
Changelog
Bug fixes and improvements
Server
- When adding a data connector agent, the agent URI can now be sourced from an environment variable (fix #9777).
- Allow ignoring the
errors
field if bothdata
anderrors
are present in the Remote Schema response. Can be configured via the serverHASURA_GRAPHQL_REMOTE_SCHEMA_PRIORITIZE_DATA
env var or theremote-schema-prioritize-data
flag. - Remove additional
internal
error response field being returned in Async Actions when dev mode is false. - Ensure Logical Models do not create empty GraphQL object types.
Console
- Also prettify the query variables on hitting the GraphiQL Prettify button (fix #9968).
- Sort the columns in the filter and sort column selectors alphabetically in the data browser (fix #9966).
Data Connectors
- Fix multi-schema issue for simple queries in Snowflake. (Cloud / Enterprise edition only)
- Fix aggregation limit for Snowflake, now aggregate limits are treated separately from row-based limits which can come through from permission-based limits. (Cloud / Enterprise edition only)
Build
- Update the ubuntu and redhat ubi base images to receive latest security updates