October 2026: Reusable Metrics, BigQuery Cost Visibility, and Less Rebuilding

Quick Answer

• Reuse more of what teams already built. Mosaic Models can link metrics across models, import existing BigQuery descriptions, and preserve more semantic logic during Apache Ossie imports.

• See what BigQuery workloads are costing. Sentinel Cost Intelligence now supports Google BigQuery, while new retention settings give administrators more control over how much monitoring history they keep.

• Do more without leaving the workflow. Dashboard authors can explore Mosaic data directly in the Dataset panel, while Report Services Document subscriptions can now be managed in Library.

Enterprise analytics gets expensive in small, repetitive ways. A modeler re-enters descriptions that are already documented in BigQuery. Revenue gets defined again because another model has a different level of detail. A semantic model moves from another platform and database-specific expressions have to be rebuilt by hand. A dashboard author switches tools just to understand the hierarchy behind an attribute.

None of these tasks are especially difficult on their own. Across a large analytics environment, every duplicate definition, manual reconciliation, and extra workflow becomes another thing to maintain, validate, and eventually fix.

October is focused on taking more of that work out.

Link Metrics Across Mosaic Models

Revenue should not become a different metric just because one model contains detailed transactions and another contains aggregated data.

Mosaic Models can now link base metrics across models with different levels of detail. Teams can define a metric once and reuse that definition across linked models instead of maintaining another copy every time the data grain changes.

For example, if a team uses one model for live daily transactions and a faster aggregated model for monthly analysis, they no longer need to build Revenue twice. The business definition stays the same even as the data grain changes.

Mosaic can suggest likely metric links during setup, and queries automatically use the linked model with the appropriate data grain. Calculations built on those linked metrics can be reused too.

The practical benefit is fewer definitions to create, validate, and maintain.

Bring Google Knowledge Catalog Context Into Mosaic Models

If table and column descriptions already exist in BigQuery, there is little value in typing them again somewhere else.

When creating a Mosaic Model from Google BigQuery, Mosaic can now import table and column descriptions from Google Knowledge Catalog.

Teams can reuse the metadata they already maintain instead of re-entering it in Mosaic, reducing manual work and another copy of business context to maintain.

Import More Complete Apache Ossie Models

Open model exchange is only useful if the model survives the move.

Real semantic models often contain database-specific functions. When a model moves between platforms using Apache Ossie, formerly Open Semantic Interchange, those expressions can become another manual cleanup project.

Mosaic can now preserve unsupported database-specific functions during Apache Ossie imports by converting them into pass-through expressions for attributes, metrics, and filters.

That means more of the original semantic logic can come across from platforms such as Databricks and Snowflake without someone rebuilding those expressions by hand.

Keep Rolling-Window Models From Growing Forever

Not every in-memory model needs to keep every historical row forever.

Mosaic Models now add scheduled delete options at the table level. Teams can remove rows matching a required filter, or delete and reload those rows with refreshed data.

A team keeping a rolling period of recent data in memory can remove older rows and load the new period without republishing the entire model.

Only the selected tables need to refresh, helping teams keep memory focused on the data people actually use.

Sentinel Cost Intelligence for Google BigQuery

In June, Sentinel Cost Intelligence connected Snowflake and Databricks spend back to the analytics assets driving it.

October extends that visibility to Google BigQuery.

Administrators can connect BigQuery billing data in Workstation and use Sentinel to monitor data-source compute spending and cost trends through a refreshed summary page.

Instead of cross-referencing a BigQuery bill with a separate analytics catalog, teams can see compute cost in the context of the governed datasets they already manage in Mosaic.

That makes the useful question less about what the cloud bill was and more about what is driving it.

Which datasets are consuming the most compute? Are refresh patterns creating unnecessary cost? Is data being kept or processed in ways that no longer make sense?

Bringing BigQuery into Sentinel gives data and analytics teams a shared view of cost and the assets behind it, making it easier to decide what to change.

Set Retention by Sentinel Module

Monitoring data has its own lifecycle.

As Sentinel captures more cost, usage, performance, and governance information, administrators need control over how much history they keep.

Workstation now lets teams configure retention periods by Sentinel module. Defaults remain in place, including 180 days for most modules and 30 days for Performance, but administrators can adjust those periods based on their own operational or compliance needs.

Sentinel then cleans up data according to the configured policy. Optimized storage and compression also help limit storage usage as monitoring history grows.

Teams can retain the monitoring history they need while keeping unnecessary storage growth under control.

Explore Mosaic Model Data While Building a Dashboard

Choosing a field by name is easy. Understanding the data behind it often takes another step.

A dashboard author may need to know which stores belong to a region, which products sit under a category, or what values actually exist before building a filter or visualization.

A new Tree View in the Dataset panel lets authors explore that information while they work.

Authors can expand Mosaic Model attributes to see their elements, move through hierarchies level by level, and search or page through attributes with large numbers of values. The experience is available in Library and the new Workstation dashboard editor.

The benefit is simple: understand the field before using it, without leaving the dashboard to investigate the model somewhere else.

That means fewer wrong turns and less back-and-forth between tools.

Manage Report Services Document Subscriptions in Library

For customers moving users from Strategy Web to Library, scheduled document delivery can be one of those long-standing workflows that still has to work before a migration is complete.

Library now supports subscriptions for Report Services Documents.

Users can schedule RSD delivery through email, file, or FTP in Excel or PDF format, with or without bookmarks. Existing RSD subscriptions created in Web can also be edited or deleted from Library.

A weekly PDF sent to regional managers can keep running. An Excel file delivered to FTP for a downstream process can keep running.

The team managing those workflows no longer has to return to Web just because the subscription was created there.

It removes another operational dependency on the older experience and brings more day-to-day BI administration into Library.

October builds on what teams already have. Linked metrics, imported metadata, and more complete semantic-model imports mean fewer definitions and expressions to recreate or reconcile. BigQuery cost visibility helps teams understand compute spend, while targeted refresh policies and configurable monitoring retention give them more control over memory use and monitoring storage. Improvements to dashboard authoring and document subscriptions also keep routine work in the same place.

Together, these improvements make mature analytics environments easier to operate without forcing teams to rebuild what already works.

Learn more about the October updates on our What's New page and Strategy software product documentation.