2026 W32

Bruin Cloud

Dashboards

• New Edit widget SQL → Open a widget's SQL in the dashboard editor, change the query, and rerun it without leaving the dashboard.
• New Conditional table formatting → Color cells with rules or gradients to make exceptions stand out. Rules can use hidden columns, so supporting calculations don't clutter the table. This is also available in DAC dashboard YAML.
• New Confidence intervals and error bars → Show uncertainty on line and bar charts, or compare estimates with a forest chart. The same visualizations are available in DAC.
• Improved Chart series editing → Rename series and choose which measures appear in a chart from the widget editor.
• New Chat on repository dashboards → Ask an agent questions about dashboards defined in repository YAML, just as you would on Cloud-created dashboards.
• Fix Line charts now leave gaps for missing values instead of drawing a misleading connection across them.

Connections and streams

• New BigQuery and Snowflake per-user OAuth → Connect agents using each user's own warehouse identity rather than a shared credential.
• New Stream sizing and destination names → Choose compute size per stream and set a custom destination table name when replicating a single table.
• New Pipeline alerts by email → Receive pipeline notifications by email alongside your existing messaging destinations.

Bruin CLI

• New Apache Spark pipelines → Build spark.sql assets with materializations, quality checks, and sensors over Spark Connect, Livy, or Thrift. Requires a CGO-enabled Bruin build; merge support depends on your target catalog.
• New Asset timeouts → Set timeout: 1h30m on an asset or default.timeout in pipeline.yml to cancel execution steps that exceed their limit.
• New Scoped incremental merges → Use materialization.incremental_predicate to restrict which destination rows a merge checks. Match the predicate to your data's update window so older matching rows aren't excluded.
• New ClickHouse upserts → Use the merge materialization strategy with declared primary keys to replace matching rows and insert new ones.
• New BigQuery partition controls → Configure partition expiration and required partition filters. Partition-aware transactional merges can limit reads to affected partitions.
• New Stripe billing analytics starter → Run bruin init stripe-bigquery for a pipeline with MRR, ARR, subscription movements, and invoice reports. Adapt the reporting definitions to your own billing conventions.
• New Cloud costs and backfills in the terminal → Explore spend with bruin cloud cost explorer and inspect grouped backfills with bruin cloud backfills list and bruin cloud backfills runs.

Bruin VS Code

• New Live dashboard preview → Open dashboard YAML and choose Preview Dashboard. Edits reload in the preview, and you can open it in a browser. The extension offers to install or update DAC when needed.
• Improved Ingestion authoring → Edit supported materialization strategies in the asset panel and choose Run with debug when troubleshooting an ingestion run.
• Improved More databases in the browser → Browse and query configured Microsoft Fabric, SQL Server, and Synapse connections from the Databases panel.
• Fix Switching assets no longer lets a late response replace the current SQL or lineage with the previous asset's results. Column lineage also handles catalog-qualified and mixed-case names more reliably.

DAC

• New Share filtered dashboards → Active filters are saved in the URL, so a colleague opens the same view you shared. Searchable dropdowns make long filter lists easier to navigate.
• New Conversion funnels → Add a funnel widget to show stage values, share of the first stage, and step-to-step conversion. Vertical and horizontal layouts are supported.

ingestr

• New New sources → Adapty, FastSpring, Payrails, OpenAI Usage, Typeform, and CleverTap.
• New SAP HANA destination → Load into SAP HANA with replace, append, merge, delete+insert, or SCD2 strategies, without installing SAP client libraries.
• Improved Partitioned custom queries and replacements → Split custom SQL extraction into windows, including full replacement loads. For replace, the selected interval becomes the entire destination dataset; use merge or delete+insert when retaining other periods.
• New Prometheus stream metrics → Add --metrics-addr to a streaming run to expose replication lag and per-table throughput at /metrics.
• Improved OneLake schema evolution → Destination tables can evolve with incoming columns instead of requiring a manual rebuild.
• Fix PostgreSQL truncate-and-insert loads now run atomically, so a failed insert doesn't leave the destination empty.

Bruin Academy

• New Analytics engineering course → Work through modeling, incremental pipelines, quality checks, and lineage in the hands-on course.