Sources
Overview
Flows can ingest data from the following sources into Onehouse tables.
| Source | Category | AWS projects | GCP projects |
|---|---|---|---|
| Amazon S3 | Files | ✅ Supported | Not supported |
| Google Cloud Storage | Files | Not supported | ✅ Supported |
| Apache Kafka | Event Stream | ✅ Supported | ✅ Supported |
| Confluent Kafka | Event Stream | ✅ Supported | ✅ Supported |
| Amazon MSK Kafka | Event Stream | ✅ Supported | Not supported |
| Amazon Kinesis (Coming soon) | Event Stream | ✅ Supported | Not supported |
| Confluent CDC | Database | ✅ Supported | ✅ Supported |
| MySQL CDC | Database | ✅ Supported | Not supported |
| Postgres CDC | Database | ✅ Supported | ✅ Supported |
| Onehouse Tables | Data Lakehouse | ✅ Supported | ✅ Supported |
Add a Source
- In the Onehouse console, open the Sources page.
- Click Add New Source.
- Follow the on-screen instructions to set up your source.
- Also review the product docs for the source to ensure you configure it properly.
- Create the Source.
Update a Source
When you change a source's connection details — for example, rotating credentials, updating broker hosts, or pointing to a new schema registry URL — apply the change in the source configuration, then perform either of the following to propagate the new configuration to the Flows that read from the source:
- Pause and Resume each Flow built on the source, or
- Restart the Managed Compute Cluster(s) running those Flows.
Both actions pick up the new source configuration without archiving the destination table or replaying historical data. Pause/Resume preserves the existing checkpoint, so each Flow continues from where it left off.
Clean & Restart archives the destination table and re-ingests from the earliest available source data. It is not the right action for routine credential rotation or other source configuration updates — use Pause/Resume instead.