v0.0.43 (2026-11-09)
- Fixes probe_integration failing with a
NullPointerExceptionwhen the request sendschecksorcustom_attributesas null, which is how Lumos sends a probe that runs every check. A null list is now treated like an omitted one
v0.0.42 (2026-11-05)
app_infonow reflects the connection’s settings: capabilities whose query is blank are left out, matchingconnected_info. Lumos no longer runs steps the connection doesn’t configure, such as Last Activity without a Get Last Activity query. Without settings,app_infostill lists every capability
v0.0.41 (2026-11-02)
- Fixes a crash (
NegativeArraySizeException) when a query returns a very large result, around 4 million rows or more. Results are now written to the agent as they are serialized instead of being built in memory first, which also lowers the connector’s peak memory use - Results are always written as UTF-8, regardless of the agent host’s default character encoding
- Adds the probe_integration capability, so a connection can be tested before it is connected. It checks that the List Accounts, List Entitlements and Find Entitlement Associations queries return data, reading at most 10 rows from each, and that every configured query can be executed
- Configured queries are checked without connecting: placeholders the capability doesn’t provide or a request may leave out, write queries that never reference the account or entitlement they change, and identical activate/deactivate or assign/unassign queries
- On MySQL, MariaDB, Postgres, SQL Server and Oracle, queries are also compiled by the database without being executed, reporting syntax errors, missing tables or columns, and read queries that don’t return a required column, with the database’s own message. On Oracle, missing privileges are reported too. DDL on Postgres and Oracle, and PL/SQL blocks, cannot be compiled without running them and are reported as not verified
- A List Entitlements query that doesn’t return is_assignable is reported as a warning, and the probe notes when none of the entitlements it read is assignable, since such entitlements can’t be requested in the App Store
- The probe confirms that sampled entitlement associations belong to real accounts by looking their accounts up with the Get Account query, when it is configured, and says so when the link could not be confirmed
v0.0.40 (2026-10-08)
- Adds a bundled CSV JDBC driver (csvjdbc 1.0.46) and a
CSVdatabase type so SQL read queries can run against CSV files on the agent host - CSV URIs accept only csvjdbc formatting options;
function.*,cryptoFilterClassNameand theclass:/classpath:path prefixes are rejected, andLUMOS_JDBC_CSV_ROOTon the agent restricts which directories CSV connections may read - Blank CSV cells are returned as null instead of empty strings
- CSV connections are read-only:
validate_credentialsrejects configured write queries and write capabilities fail before connecting - Boolean fields accept
true/false(any case) as well as1/0for all database types
v0.0.39 (2026-09-02)
- Fixes bug where Create Account Return ID query was skipped for DDL statements
- DDL writes and the Return ID query now run in a transaction on engines that support transactional DDL statements (Postgres and DB2)
v0.0.38 (2026-08-13)
- Updates bundled third-party libraries to address reported security vulnerabilities: Jackson to 2.18.8, json-smart to 2.5.2, and the Databricks JDBC driver to 3.4.2
v0.0.37 (2026-06-24)
- Adds support for connecting to SQL Server using Windows integrated authentication, so customers can use the OnPrem Agent service account instead of storing SQL username and password credentials
v0.0.36 (2026-05-19)
- Adds the update_account capability to the connector, allowing updates to account records using custom SQL queries with support for string, boolean, and number custom attributes
v0.0.35 (2026-04-16)
- Adds the update_account capability to the connector, allowing updates to account records using custom SQL queries with support for string, boolean, and number custom attributes
v0.0.34 (2026-04-10)
- Add alternative aliases integ_specific_entitlement_id and integ_specific_resource_id for users of Oracle DB less than version 12.2
v0.0.33 (2025-11-13)
- Update app_info to properly serialize x-semantic types
v0.0.32 (2025-10-24)
- Update Databricks to use PAT auth
v0.0.31 (2025-10-16)
- Adds support for Databricks
v0.0.30 (2025-10-07)
- Add third party licenses to distribution
v0.0.29 (2025-09-30)
- Adds support for MariaDB databases
v0.0.28 (2025-09-03)
- Adds the list_updated_accounts capability to the connector, which can be used to perform incremental syncs of data
v0.0.27 (2025-08-04)
- Fixes an issue with app_info formatting
v0.0.26 (2025-07-21)
- Adds the get_account capability to the connector
v0.0.25 (2025-07-11)
- Fixes an issue with app_info formatting
v0.0.24 (2025-05-27)
- Fixes an issue where app_info sometimes would not work with agent
v0.0.23 (2025-05-09)
- Fixes an issue where connected capabilities may be incorrectly reported when queries are set to an empty string
v0.0.22 (2025-05-08)
- Adds app_info endpoint
v0.0.21 (2025-05-08)
- Updates processing of queries for Oracle DBs to correctly handle trailing semicolons
v0.0.20 (2025-04-28)
- Adds the Snowflake JDBC driver
v0.0.19 (2025-04-16)
- Adds a new optional field “Create Account Return ID” which can be filled out to return the id of the account created by the “Create Account”. This query will be run in the same transaction as the “Create Account” query and is necessary to support Lumos Onboarding flows.
- Adds support for returning the field “id” from “Create Account” if “Create Account” can be executed as a single statement (i.e. executing a store procedure). This is an alternative to using the “Create Account Return ID” setting.
v0.0.18 (2025-04-10)
- Adds a better error message when the JDBC URI is incorrect
- Automatically remove a trailing semicolon from Oracle DB queries
- Adds instructions with a link to the developer docs to JDBC settings