This message was deleted.

Yes. You can ingest into two separate datasources and each can have its separate retention settings. What's the use case? The community may be able to provide better guidance if you describe what you are trying to do. Is this streaming or batch data?

We provide a metric service to our internal teams. And we have many use cases, some use cases could need streaming and some others batch.

We see implementations where a single stream is consumed into both a raw events and an rollup data source with separate streaming jobs. As a note, streaming ingestions should also setup auto-compaction. With two separate datasources, you can then set the retention period for each. Depending on you use case, you may be able to pull off a single datasource and use compaction to change the granularity of the rollup and the segments as the data ages. You can also do this with batch ingestions & manual compaction using SQL based ingestion jobs. Where you might ingest raw events and then use a reindex to rollup older intervals.

Thank you Sergio, it's clear for now