Skip to main content
Version: Tenzir v4.5


This page defines central terms in the Tenzir ecosystem.

missing term?

If you are missing a term, please open a GitHub Discussion or ping us in our Discord chat.


Web user interface to access platform at

The app is a web application that partially runs in the user's browser. It is written in Svelte.


Maintains partition ownership and metadata.

The catalog is a component in the node that owns the partitions, keeps metadata about them, and maintains a set of sparse secondary indexes to identify relevant partitions for a given query. It offers a transactional interface for adding and removing partitions.


Manages chunks of raw bytes by interacting with a resource.

A connector is either a loader that acquires bytes from a resource, or a saver that sends bytes to a resource. Example connectors are file, kafka, and nic.


Translates between bytes and events.

A format is either a parser that converts bytes to events, or a printer that converts events to bytes. Example formats are json, cef, and pcap.


Optional data structures for accelerating historical queries.

Tenzir has sparse indexes. Sparse indexes live in memory and point to partitions.


A host for pipelines and storage reachable over the network.

The tenzir-node binary starts a node in a dedicated server process that listens on TCP port 5158.


Runtime statistics about pipeline execution.


The building block of a pipeline.

An operator is a source, transformation, or sink.


The horizontal scaling unit of the storage attached to a node.

A partition contains the raw data and optionally a set of indexes. Supported formats are Parquet or Feather.


Combines a set of operators into a dataflow graph.


Control plane for nodes and pipelines, accessible through app at


A named record type describing the top-level structure of a data frame.



An operator consuming input, without producing any output.


An operator producing output, without consuming any input.


An acronym for Tenzir Query Language.

TQL is the language in which users write pipelines.


An operator consuming both input and producing output.