What problem does it solve? Migrating Azure Event Hubs streaming ingestion to Google Cloud traditionally requires a costly Dataflow fleet of GCE workers. This Skill guides the replacement of that fleet with Pub/Sub Import Topics and BigQuery direct subscriptions writing to BigLake Apache Iceberg tables, cutting ingestion costs by roughly 95% while preserving legacy metadata column contracts. ## Core Features & Use Cases - OIDC Workload Identity Federation: Configure Azure Entra ID App Registrations, federated credentials, and GCP IAM bindings so Pub/Sub can pull from Event Hubs without static SAS connection strings. - SMT JavaScript UDF Transforms: Inject metadata columns (_meta_source_signature, _meta_namespace, _meta_enqueued_time, _meta_partition_id, _meta_sequence_number) so BigQuery rows match legacy Dataflow output contracts. - Terraform Modules & DLQ Architecture: Provision import topics, BigQuery subscriptions with use_table_schema, and a unified regional dead letter queue with correct service agent IAM. - Use Case: A data engineer needs to move 102 Event Hubs pipelines off a 258-VM Dataflow fleet in southamerica-east1; this Skill walks through the 6-phase lifecycle from Azure RBAC to parity audit and cutover. ## Quick Start Ask the assistant to provision a Pub/Sub Import Topic and BigQuery Iceberg subscription for a specific Azure Event Hub namespace using the 6-phase migration lifecycle.