Strategy ONE

Dashboard and Bot Consumption and Authoring Failover

MicroStrategy ONE (December 2024) introduces seamless session failover for dashboard and bot consumption and authoring, when scaling in Intelligence Server replicas or if the environment experiences an unexpected termination in containerized environments. The failover ensures minimal data loss and user impact by automatically backing up sessions on the backend and restoring sessions when users continue operations.

Only sessions from modern clients, including Library Web, Library Mobile, Workstation, and Hyper clients, are supported.

Failover is enabled by default in all Google Cloud Platform and Azure MicroStrategy containerized offerings.

Use Cases

Failover occurs in the following scenarios:

  1. Scaling in: When an Intelligence Server pod shuts down during scaling, users that are interacting with dashboards or bots do not experience interruptions.

  2. Maintenance or crash scenarios: If an Intelligence Server pod goes down unexpectedly (for example, if it is killed by Kubernetes), users that are interacting with dashboards or bots do not experience interruptions.

Considerations

Failover provides a best-effort recovery solution with a backup frequency of five seconds. However, users may encounter errors such as Message/Resource Not Found if an Intelligence Server shuts down and the corresponding session data is not backed up.

Session failover in the following scenarios is not supported in the current release. Users will encounter errors when they continue operations after the Intelligence Server handling their connection fails.

  • Create a new bot and the Intelligence Server is down before you save the bot.

  • Create a new document from a cube or report in Workstation and the Intelligence Server is down before you save the document.

  • In data wrangling, before the user saves the end result and the Intelligence Server is down.

  • Edit an unsaved cube or an existing non-managed cube inside the dashboard and select Keep Change Local and the Intelligence Server goes down before you save the cube.

  • In a dashboard, if the Intelligence Server handling the initial connection is down, and the user uploads an image with a new, restored session, then clicks Undo and Redo.

  • In a dashboard, if you perform the drill operation twice and click Sort, and the Intelligence Server is down before you click Undo.

  • Report and HyperIntelligence Card consumption and authoring.