Hi Metaflowers, How can I cleanup old runs in the metadata-service/artifact store? As for the artifacts I could delete them from the S3 bucket. However is there a canonical approach to this?

@gentle-insurance-89626 - are you concerned about the storage in the db or in s3? within the db, you can drop old rows however in s3 - we use content addressed storage so dropping data may affect future executions. The content addressed store is per flow - so you can definitely drop data for the entire flow without affecting other flows (only future executions of the same flow might be affected)

I'm concerned about the storage requirements of the artifacts. In my usecase I generate a lot of intermediate artifacts that are adding up now.

the data in s3 is heavily compressed and only a single copy if ever stored for a flow - that limits the pace at which your storage costs grow quite significantly.

So is there a canonical way to delete old flows, via API or CLI?

there isn't currently

OK thanks!