1. Documentation
2. 10.3 (Latest) 10.2 10.1 10.0 9.10 9.9 9.8 9.7 9.6 9.5 9.4 9.3 9.2 9.1 9.0 8.2 8.1 8.0 7.6
3. Hub
4. Releases
5. 10.0.0

What's New in RapidMiner AI Hub 10.0.0?

Released: 3rd November 2022

The following describes the enhancements and bug fixes in AI Hub 10.0.0:

Enhancements

• Upgrade all dependencies and Spring Boot to version 2
• Move from ActiveMQ to ActiveMQ Artemis (compatible with 2.26.0)
• Use Keycloak as Identity Provider
• Add migration path from AI Hub/Server 9
• Dynamically assign ENCRYPTION_SECRET and write into home directory if not present
  • Encryption secret will be generated dynamically once on startup and stored in encryption.secret-location
  • It is possible to overwrite the encryption secret by setting the property encryption.secret
• Use Single Page Application for frontend
• Use Postgres as only option for database
• Use proper database migrator tool for bootstrapping required schemas entirely
• Enrich Jobs API
  • Add ID of the schedule that scheduled a job
  • Jobs can be filtered by scheduleId
• License can be set via license.license and will be stored in the license.location
• Add configurable cleanup scheduler for the RTS
  • Can be enabled with scoring-agent.execution.cleanup.cron-expression
  • A cron expression has to be definedscoring-agent.execution.cleanup.enabled
  • The scheduler will wait for jobs to finish
    • A timeout can be set via scoring-agent.execution.cleanup.timeout
    • The waiting interval for running jobs can be changed via scoring-agent.execution.cleanup.waitBetween
• Add pause and resume all schedules functionality (state change)
• Add ability to toggle scheduler into standby without changing state of schedules
• Add database indexes
  • jobservice_job: created_at
  • a_jobservice_job: created_at
  • jobservice_job: stop_requested = 1
  • a_jobservice_job: stop_requested = 1
  • jobservice_job_agent: state
  • jobservice_job_agent: last_status_at
• Switch to REST endpoint for retrieving AI Hub instance information
• Add REST endpoint for log retrieving
• Add new Spring Security classes for authorization in the Scoring Agent
• Add migration to the Scoring Agent
• Change units from minutes to seconds for jobservice.scheduled.archive.maxAge and jobservice.scheduled.archive.jobCleanup.maxAge
• Remove jupyterEnabled and grafanaEnabled from endpoints and UI which is now handled by the platform setup
• Disabling the automatic snapshot creation after a job has been executed inside a Job Container can now be achieved by setting the Job Container's property jobcontainer.remote.versioned.auto-commit-strategy to NONE
• Change host of rm-juypterhub-db-svc to jupyterhub-db in the sample-dev Project
• Respect X-Forwarded-.. HTTP headers for Project URL generation natively

Fixed

• Add missing internal locking of Job Agent scheduled tasks
• Change execution mode of Scoring Agent to SCORING_AGENT instead of JOB_CONTAINER
• Strip common 80 and 443 ports from LFS server responses