| Agent |
A saved assistant with instructions, a model and a fixed set of tools. Other systems can call it after a reviewer approves a version. |
| Approval |
A confirmation the assistant asks for before it changes something. |
| Batch / Streaming |
The two schedule types in the Scheduler. Streaming (the default) runs once; Batch repeats. |
| Catalog |
The searchable list of datasets, ETL jobs and warehouses. |
| Compute |
The memory and processor size given to a runtime. |
| Connection |
A saved link to an outside system. Code uses it by name and holds no credentials. |
| Contract tab |
A Catalog tab that shows how to call a dataset or job from outside Groundzero. |
| Dashboard |
A page of visualizations in tabs, with filters. You can publish it and import it into a workspace. |
| Dataset |
A saved SQL or code query whose result charts and dashboards use. |
| ETL job |
Python code in a project that reads, changes and writes data. |
| Lakehouse / Warehouse |
Groundzero storage for governed tables, organised in databases and tables. |
| Lineage |
A map of how data flows between tables, jobs and datasets. |
| MCP |
Model Context Protocol, a standard way for tools and assistants to work with Groundzero. |
| Ontology |
Shared business definitions (entities, metrics, terms) that the assistant can use as context. |
| Project |
A DE/ML container for jobs, notebooks, models and agents. |
| Runtime (RTE) |
The environment that runs your code: Standard, DuckDB or Spark, and their variants. |
| Session |
A running notebook. You start it, work in it and stop it. |
| Site |
Your organisation's Groundzero, with its own logo and theme. |
| System variable |
A placeholder such as ${JOBNM} that Groundzero fills in before your code runs. |
| Visualization |
A saved chart or table on a dataset. |
| Workspace |
A named folder that holds imported (published) dashboards. |
