Pipeline run tracking

Every analysis run in Seqera carries a status as it moves from queued to complete or failed. A Beam agent watches for status changes, reads the run record, and applies the rule your team approved: on a clean finish it notes the outputs and notifies the requester, and on a failure it records the error and opens a ticket. Runs that stall midway, or finish with warnings your rule cannot classify, go to a scientist with the log attached, so a person decides whether to rerun the job or investigate the underlying cause.

Dataset intake

New datasets registered in Seqera need naming, tagging, and linking to the right workspace before analysis begins. When a dataset appears, the agent reads its metadata, checks it against your naming and permission rules, and updates the tags or workspace assignment where the connector allows. It notifies the owner once the dataset is ready to use. Datasets missing required fields, or arriving with sizes or formats that fall outside your approved list, route to a data steward, who confirms the details before the agent files them for downstream pipelines.

Compute environment checks

Seqera ties each run to a compute environment that defines where the work executes. A Beam agent reviews environment records on a schedule, reads which ones are active and how they map to workspaces, and applies your rule for keeping them tidy: idle or duplicate entries get flagged, and the owner is notified with the details. Environments tied to running jobs, or ones whose settings your rule does not cover, stay untouched and go to an administrator, who decides whether to retire, edit, or leave the configuration in place.