Lakeflow Connect added managed marketing connectors and Lakeflow Jobs released task output in for_each loops.

Databricks expanded Lakeflow Connect with native managed connectors for ad and marketing platforms, including Salesforce, HubSpot, Google Ads, and Meta Ads, landing raw ingestion data into Unity Catalog [29]. Accompanying guides demonstrated how to model these feeds for marketing analytics [28]. Within Lakeflow Jobs, support for passing task outputs inside for_each task loops entered public beta [3], while Databricks delivered updates to Lakeflow Designer [14].

Teams adopting Lakeflow Connect shared early production experiences and raised infrastructure questions [19]. Users deploying the SQL Server change data capture gateway asked how to select specific virtual machine types [10], handle elastic compute requirements [1], and avoid connection saturation on source transactional databases [22]. Practitioners also shared architectural patterns for managing upstream schema drift and executing historical backfills alongside ongoing change streams [6].

Lakeflow Jobs sparked discussion around production orchestration maturity. Teams evaluated replacing legacy Airflow schedules with data-aware triggers in Lakeflow [16, 26], though engineers debated whether the platform provided adequate tooling for service level objectives and SLA monitoring across large fleets [15, 25]. Community members also worked through practical job configurations, such as implementing conditional task retries [8] and enabling predictive optimization for streaming tables in declarative pipelines [18].