Every run is logged with per-step inputs, outputs, durations, and errors. Filter by status, record, trigger, and date. This is your debugging surface and your audit trail.
The run list
Automation → [workflow] → Runs: status (completed/failed/partial), trigger source, target record, duration, version. Filters make "show me failures on enterprise accounts this week" a ten-second question.
Per-run detail
Click a run for the step-by-step: each step's input payload, output, duration, and — for branches — which condition matched. Agentic steps include the reasoning summary.
Failure triage
Failed steps show the error class (auth, rate-limit, mapping, null-field) and a retry button. Bulk-retry handles transient outages: filter failures by time window, retry all.
Retention and export
Run history follows workspace retention (default 12 months), exports to CSV, and streams to the warehouse export for teams that dashboard automation health externally.
