This section covers how data is structured and ingested into Rerun.
Recordings
How recordings are identified and grouped by application ID
Entities and Components
How entities and components form Rerun's data model
The Entity Path Hierarchy
How entity paths organize data into a hierarchy
Transforms & Coordinate Frames
Coordinate frames and how transforms relate them
Events and Timelines
Events, timelines, and how Rerun tracks time
Static data
Logging data that holds across every timeline
Chunks
The Arrow-backed storage unit Rerun uses internally
RRD format
How chunks are framed and serialized in .rrd files
Chunk Processing API
Pipelines for ingesting, transforming, and converting chunks (experimental)
Query semantics & partial updates
How Rerun resolves queries at any point in time
MCAP files
How Rerun reads and interprets MCAP files
Component Batches
Efficiently log collections by batching component values
Sinks
The destinations logged data can be sent to
Video
How Rerun handles video streams and codecs