databricks-zerobus-ingest

Ingest records into Databricks Delta tables via the Zerobus gRPC API.

Updated Apr 18, 2026
One-click install
npx skills add https://github.com/aaronachermann/PolentaEncoders --skill databricks-zerobus-ingest-aaronachermann
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: databricks-zerobus-ingest
Source: https://github.com/aaronachermann/PolentaEncoders/tree/main/.github/skills/databricks-zerobus-ingest
Command: npx skills add https://github.com/aaronachermann/PolentaEncoders --skill databricks-zerobus-ingest-aaronachermann

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Zerobus Ingest removes the need for external message buses by enabling producers to write records directly into Databricks managed Delta tables with durability acknowledgments, reducing operational complexity and end-to-end latency for near real-time ingestion.

Core Features & Use Cases

  • Direct gRPC ingestion: Build clients that talk to the Zerobus gRPC API to stream records into Unity Catalog managed Delta tables.
  • Multi-language SDKs: Examples and patterns for Python, Java, Go, TypeScript/Node.js, and Rust with both JSON and Protobuf support.
  • Schema and safety: Generate Protobuf from Unity Catalog schemas, enforce message size and column limits, and require explicit table-level grants for secure writes.
  • Operational patterns: ACK tracking, callback-based confirmation, retry with exponential backoff, stream reinitialization, flush semantics, and throughput scaling by opening multiple streams.

Quick Start

Create a Zerobus stream and ingest a sample JSON record into your managed Delta table using the installed SDK and service principal credentials.

Frequently Asked Questions about databricks-zerobus-ingest

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I stream data directly into Databricks Delta tables without an external message bus?

You can stream data directly into Databricks Delta tables by using the Zerobus gRPC API for record-level ingestion, removing the need for external message buses while reducing operational complexity and latency.

Can I ingest JSON and Protobuf records into Unity Catalog managed tables using gRPC?

Yes, you can ingest events using gRPC with either Protobuf or JSON serialization into Unity Catalog managed tables, generating Protobuf schemas directly from Unity Catalog to enforce message constraints.

What programming languages are supported for building Databricks event ingestion SDK clients?

You can build Databricks event ingestion SDK clients in Python, Java, Go, TypeScript, or Rust, utilizing provided patterns for near real-time telemetry, IoT, or application-event streaming.

Do I need service principal credentials to write records to Delta tables via gRPC?

Yes, writing records via gRPC requires service principal credentials with explicit MODIFY and SELECT grants to ensure secure writes to the managed Delta tables.

How do I handle retry logic and durability acknowledgments for streaming data into Databricks?

You handle retry logic and durability by using ACK-based tracking, callback-based confirmations, and retrying with exponential backoff, along with stream reinitialization to ensure reliable ingestion.

How can I scale throughput for near real-time telemetry ingestion into Databricks?

You can scale throughput for telemetry ingestion by opening multiple streams, while enforcing per-stream throughput limits, message size constraints, and schema column limits to maintain stability.