Skip to main content

Celonis Product Documentation

Installing and updating the on-premise JDBC extractor


Any references to third-party products or services do not constitute Celonis Product Documentation nor do they create any contractual obligations. This material is for informational purposes only and is subject to change without notice.

Celonis does not warrant the availability, accuracy, reliability, completeness, or usefulness of any information regarding the subject of third-party services or systems.

Our on-prem agents and extractors are available from the Download Portal in Celonis. Go to Admin and Settings > Download Portal to get and update them, and to see the full release notes. You'll need to be an Admin or an Analyst with the permission MANAGE DOWNLOAD PORTAL for access.

For the Skill agent package, contact

We've recently upgraded to Java 17. From version 2.80.1 of the JDBC extractor, you're required to upgrade your Java version to at least Java 17 to ensure compatibility and leverage the latest enhancements and security features. With version 2.82.0, we’ve removed some additional steps from the upgrade process, so use this or a later version of the extractor package.

Make sure that all the other requirements of the extractor server are fulfilled. For the complete requirements, see System requirements of an on-premise extractor server.

Follow these steps to upgrade your on-premise JDBC extractor package, and upgrade to Java 17 or higher if you need to. Java 17 is the minimum required version, and we’ve carried out our testing with it.

  1. If you're upgrading from a version of the JDBC extractor earlier than 2.82.0, and need to upgrade your Java version, first install Java 17 (or higher) in your local environment. System requirements of an on-premise extractor server has the exact Java package requirements. Update your environment settings to use Java 17 (or the higher version you installed) as the default Java version.

  2. Download the latest extractor package for the JDBC extractor from Admin and Settings > Download Portal in Celonis. You'll need to be an Admin or an Analyst with the permission MANAGE DOWNLOAD PORTAL for access.

  3. If you've customized the JDBC extractor's YAML configuration file application-local.yml, proxy configuration file proxy.yml, or XML configuration file CelonisJdbcExtractor.xml, save a copy of these files from the existing extractor directory.

  4. In a sandbox environment, create a new directory, and install the new version of the extractor package into it. The package creates subdirectories and extracts the files from the jar file. (It's important to do this in a new directory, rather than replacing the jar file in the existing directory, so that the new package can fully validate that it's been correctly installed.)

  5. If you're upgrading from a version of the JDBC extractor earlier than 2.82.0, transfer your customizations for the file CelonisJdbcExtractor.xml, to the CelonisJdbcExtractor.xml file supplied with the new version of the extractor package. You need to use the updated version of this file from 2.82.0 or later.

  6. If you've already upgraded to 2.82.0 or beyond, replace the CelonisJdbcExtractor.xml file in your new directory with your customized version.

  7. Replace the YAML configuration file application-local.yml, and the proxy configuration file proxy.yml, in your new directory, with your customized versions if applicable. (You don't need to retain the updated versions of these files.)

  8. Start the new version of the extractor following the instructions in "Step D: Run The Extractor" in How do I set up an on-premise Extractor?

  9. Verify in your sandbox environment that the extractor is working correctly.

  10. When you're happy with the installation, install the new version of the extractor in your production environment in the same way.

  11. Stop and uninstall the older version of the extractor when the new service is working correctly in production.

Change history





  • Added support for Optimizer hints while extracting data from Oracle databases.

  • The Windows installer executable is now signed with a Celonis certificate.

  • Unified the three metadata retrieval approaches so that they all return consistent information.

  • Upgraded to newer versions of the logback and json libraries.



  • Fixed Databricks extraction issues when using the default catalog and default database.

  • Fixed some security vulnerabilities.



  • For driver metadata, fixed Microsoft SQL server extraction issue when table has clustered columnstore index.

  • Optimized memory allocations during extractions.



  • Beta release of Oracle Smart Extraction, which parallelizes extractions of larger Oracle tables to reduce data extraction times. The feature is shipped disabled. If you want to try it out, we recommend that you do so in a sandbox environment. To get Oracle Smart Extraction enabled, talk to your Celonis point of contact or create a support ticket.

  • Fix for an SAP HANA filter parser error when concatenation is used.



  • Fixed deviations for dates earlier than 1900 due to timezone changes.

  • Fixed an issue with the Snowflake driver for a new installation on Microsoft Windows.

  • SQL ID will now be logged for Oracle if debug mode is enabled.



  • Enabled STRING to DATETIME conversion for BigQuery and Trino.

  • For Oracle, we’ve improved the query for the INFORMATION_SCHEMA metadata source.

  • Quotes in the filter statement are now recognized.

  • Fixed some security vulnerabilities.

  • Fixed an issue for the custom BigQuery driver where classes were not loaded in the correct order.

  • Fixed an issue for uplinked extractors using a proxy configuration.



  • Upgraded the JDBC extractor’s internal libraries. If you’re linking the BigQuery driver, you’ll need to exclude all SLF4J .jar files from the driver package.

  • On Microsoft Windows, we’ve changed the JDBC extractor’s dependency from the Microsoft Visual C++ 2010 Redistributable Package to the Microsoft Visual C++ 2015-2019 Redistributable Package (x64). Install that package when you install this version of the JDBC extractor.

  • Wildcards in Snowflake metadata calls are now escaped to improve load.



  • Added support for Oracle CLOB (Character Large Object) and NCLOB (National Character Large Object) data types.



  • Extractions that hang can be resumed from the last table, instead of restarting them.

  • Upgraded the driver for Snowflake to version 3.13.33.

  • Upgraded the driver for Athena to version 2.0.36.

  • Upgraded the driver for IBM DB2 to version 11.5.

  • Implemented TO_DATE functionality for Oracle filters.



and 2.80.1

  • Upgraded the JDBC extractor to Java 17.

  • From version 2.80.1, you're required to upgrade your Java version to at least Java 17 to ensure compatibility and leverage the latest enhancements and security features.

  • With version 2.82.0, we’ve removed some additional steps from the upgrade process, so use this or a later version of the extractor package.



  • Upgraded MSSQL-JDBC Driver to latest version.

  • Set trustServerCertificate=true and encrypt=false by default in case they are not set in the additional properties field (required by driver upgrade).



  • Upgraded MySQL driver to latest version.



  • Fixed security vulnerabilities.



  • Added support for extractions from Databricks.

  • Oracle DB: Fix for scenarios where the driver metadata was used, even if Information Schema is selected.

  • Improved clean-up of changelog tables for real-time extractions by doing the clean-up in chunks.



  • Improved metadata query for Oracle databases.



  • Added support for Analytical Views for SAP HANA.

  • Fixed the feature to clear the metadata cache.



  • Added support for Analytical Views for SAP HANA.



  • Fixed Java-based vulnerabilities.



  • Added support for key pair authentication for Snowflake.



  • Extraction of synonyms for Oracle databases.



  • Extended the logging messages.



  • Enabled Materialized view for the Postgres database.

  • Added support for Attribute(Joined) views for SAP HANA.



  • BigQuery Get Tables from Additional Projects.

  • BigQuery ADC authentication (hosted in GCP).

  • Fixed test connection issue for BigQuery.

  • Fixed input box for BigQuery data connection form.



Minor improvements and fixes.



  • Added SHA-256 and SHA-512 support.

  • Fixed the order of delete and insert executed at the same time.

2.57.0 (2022-05-05)

  • For Google BigQuery, fixed duplicate records caused by a LIMIT/OFFSET in SELECT queries without ORDER BY (primary key) clause not guaranteeing proper pagination.

  • Added validation to inform the user with an error message if a primary key is not selected.

  • Fixed NullPointerException in uploading results leading to duplicated push jobs.

  • Extended invalidate cache also for real-time integration column selection.

2.56.0 (2022-04-21)

  • Added initialization for JDBC real-time via Replication Cockpit.

  • Fixed out of memory error due to unlimited threads.

  • Fixed column order changes after deselecting some of the columns in JDBC.

2.55.0 (2022-04-07)

  • Added a database connection timeout setting in the UI. This overwrites the local timeout in application-local.yml in case of uplinked connections.

2.54.0 (2022-03-24)

  • Added an authentication option SERVICE_ACCOUNT_AUTHENTICATION for Google BigQuery database connection. Inputs are the service authentication account email ID and the service account key file.

  • Improved logging for JDBC extraction in DEBUG extraction mode.

  • Added support for Vertica database type.

  • Performance improvements for JDBC data extractions on the upload mechanism.

2.51.0 (2022-02-11)

  • Removed the option to include a changelog time stamp in JDBC real-time extractions, and made it the default, to support real-time transformations.

2.50.0 (2022-01-21)

  • Extended duplicate removal for all database types by adding properties in application-local.yml for uplink database connections. To enable this, add the following to the application-local.yml file:

       enabled: true   
         strategy: CLOUD

2.49.0 (2022-01-13)

  • Logical change in reading change log tables in the JDBC real-time scenario to improve performance.

  • Fixed change in metadata source not being consistently reflected in the metadata query.