Inspect and control runs

Toolang records execution history, allowing you to inspect past runs, monitor ongoing work, and control active runs.

Inspect runs

List recorded threads, runs, and controls:

too eve inspect threads
too eve inspect runs
too eve inspect controls

Use an ID from the listings to inspect execution details:

too eve inspect THREAD runs
too eve inspect RUN steps
too eve inspect RUN tree
too eve inspect RUN output
too eve inspect STEP call

These commands let you explore a thread's runs, the steps within a run, its execution tree, output, and individual calls.

call applies to model, tool, run, par, and loop steps. For value steps such as let, inspect the step directly:

too eve inspect STEP

Add --json for machine-readable output. Inspection reads stored history and works without a running agent server.

For a standalone script, use @ to inspect its latest retained instance:

too helpers.too@ inspect runs

Use helpers.too@CODE to select a specific instance. While it is running, too ps shows its instance code.

Control active runs

Steer an active run with additional instructions or request cancellation:

too eve steer RUN "Focus on the smallest safe change."
too eve cancel RUN

These commands require the owning server to be running. A thread ID can also be used to target its active run.

Cancellation does not delete recorded history.

Monitor runs

Use top to monitor active runs, recent completions, and resource usage:

too top
too eve top --period 4h --recent 15m
too eve top --once
too eve top --json

top combines current execution, recent completions, and rolling usage statistics. Without an agent name, it uses the Toolang root.

Use --once for a single snapshot or --json for machine-readable output. For process and Hub status, use too ps.