ContextQMD
Back to Datahub

DataHub Actions

datahub-actions/README.md

1.7.02.1 KB
Original Source
<!-- PyPI long description. Keep concise, feature-discovery-first. -->

DataHub Actions

Event-driven automation for DataHub — react to metadata changes in real time and trigger workflows, notifications, or custom logic.

What you can do

  • Listen to metadata change events as they happen (entity updates, tag additions, ownership changes, etc.)
  • Filter events to only act on what matters to you
  • Chain transformers to enrich or reshape events before they reach your action
  • Build custom actions — call APIs, send Slack alerts, trigger pipelines, sync to external systems
  • Run multiple pipelines in a single process

Installation

bash
pip install acryl-datahub acryl-datahub-actions
datahub actions version

Quickstart

Define an action pipeline in YAML and run it:

yaml
# my_action.yml
name: tag_watcher

source:
  type: kafka
  config:
    connection:
      bootstrap: localhost:9092
      schema_registry_url: http://localhost:8081

filter:
  event_type: EntityChangeEvent_v1
  event:
    category: TAG
    operation: ADD

action:
  type: hello_world
bash
datahub actions -c my_action.yml

Key concepts

ConceptDescription
SourceWhere events come from (currently: Kafka)
FilterNarrow down which events trigger the action
TransformerOptionally reshape or enrich events
ActionWhat to do — call an API, send a message, run code

Supported events

  • EntityChangeEvent_v1 — fired when an entity's metadata changes (tags, owners, terms, etc.)
  • MetadataChangeLogEvent_v1 — low-level log of every aspect write