Connectors

A connector moves data between external systems and PADAS streams: sources ingest into a stream; sinks read from a stream and deliver outward. Lifecycle is the same for every class (stopped → starting → running → stopping) and is driven from the REST API Reference and PADAS UI.

For the platform model, see Core concepts — Connectors and Architecture — Connectors (I/O boundary).

Pages in this section

Topic	Use it when
Sources	You configure ingest under Sources (/sources).
Sinks	You configure delivery under Sinks (/sinks).
Connector types	Class-specific UI, config, and Required fields on each type page below.

Every connector row shares top-level fields (name, class, stream, enabled, …) on the Sources/Sinks screens; the large class-specific shape lives under config and differs by class. Mandatory name, class, stream, and required config keys for each class are documented under Required fields on the matching connector type page.

Connector classes

Registry exports (and the UI Advanced Settings forms) use a wide config object: many keys exist for all classes, but each class only uses the subset that applies. Treat the UI and an exported registry bundle as the source of truth for your build.

Class (class)	Typical role	Reference
file	Mostly source (tail / batch read); sink where supported	File connector
http	Source and sink	HTTP connector
kafka	Source and sink	Kafka connector
object_storage	Sink (S3-compatible archives)	S3 / object storage connector
padas	Source and sink (node-to-node)	PADAS connector
splunk	Source and/or sink (Splunk HEC–oriented)	Splunk connector
syslog	Source and sink	Syslog connector

Example registry snippets with filled config shapes live in padas-registry-quickstart.json (syslog source and sink).

Related pages

• Architecture — Data flow — how connectors fit the Core data path
• Streams (advanced) — stream ids connectors read and write
• Tasks — stream consumption in processing and detection
• Pipelines — wiring sources, tasks, and sinks