Running & debugging
Run configuration, breakpoints, run pages, logs and output, and controlling live runs.
10 articles
- Run a graphSupply the input, clear validation, start the run, and find it afterwards.
- Run configuration and presetsPer-node arguments for a run, saved as named presets you can duplicate, export, and seed from the last run.
- Debug with breakpointsPause execution at a node, step, inspect agent memory, and manage breakpoints across graphs.
- Read run history and tracesThe runs list, per-run traces and logs, and the run pages hanging off a graph.
- Live runs, suspension, and resumingWhere a currently-executing run shows up, why its history page can 404, and how suspended runs continue.
- Read a run's pageWhat each panel of a run's page shows — summary, controls, path, logs, output, and the execution trace.
- Run logs and outputFilter a run's log by level and category, find text in it, and read what the graph printed to stdout and stderr.
- Follow a run's path through the graphsSee every level a run entered — sub-graphs included — and the nodes it executed in each.
- Pause, stop, and resume a runControl a live run from its page, resume a paused or suspended one, or reopen it in the debugger.
- Running a bot on a scheduleGive a graph a calendar so it runs on its own — once at a set time, daily, every other day, weekly, or any combination.