Skip to content

Runs

Runs overview

The Runs screen (sidebar → Activity → Runs) is the run-centric view of your automations: one row per run — the root event of a trigger and every chained event it fans out into, collapsed into a single row. Click a run to open its trace, which flies in over the list with a back row.

Runs table

ColumnDescription
FlowThe flow the run executed
SourceThe source that fired the run — the flow’s schedule source for cron fires, its MCP source for agent tool calls, Manual for grid run-now, or the flow’s first source for older runs
TriggerHow the flow fired — Schedule, Webhook, File watch, MQTT — with the specific trigger’s identity next to it (Schedule · 0 9 * * *, File watch · /data/in), Manual for runs fired by the flows grid’s run-now action, or — when the run wasn’t flow-triggered
StatusThe worst status across the run’s events — RUNNING while any branch runs, then the worst terminal state (TIMED_OUT > ERROR > WARNING > SUCCESS); superseded events don’t count
DurationThe sum of the run’s tasks’ execution times (— for runs that predate duration tracking)
EventsHow many events the run’s fan-out chain produced
StartedWhen the run’s first event was created

Run trace

Clicking a run opens its trace as a fly-in detail:

  • Branches — one expandable row per event of the run, with the flow, status, duration, the sink that ran it, the script-declared outcome (code and label, when the script set one), the branch’s progress (the script’s latest progress marker, updating live), and a raw output link that opens the full task output in a new tab.
  • Module spans — the runner wraps every module execution automatically, so runs without script markers still get a waterfall.
  • Steps — scripts narrate their own work with step markers (see Actions & Libraries); a step’s start reads ▶ and its finish ■ with the step’s duration and result.
  • Lines — expand a node to read its lines; structured log markers render in their level’s color (italic), and raw output lines show the script’s message without the runner’s own log header.

The waterfall grows live while the run executes; the back row returns to the list.

The search all records link in the header opens the global log search — the flat table of every structured log record across runs. Clicking a record opens that record’s run trace.