Automation

Run History

Every change made through automation is recorded as a run. The run history is your audit trail: who started it, with which parameters, on which servers, what each server returned and how it ended.

What each run records

The playbook, template or command, the values used, whether it was a dry run, the rollout strategy and the person who started it.

For each server: the preflight results, each step with its outcome and attempts, the full output in the order it was printed, the exit code and the duration.

Follow a run while it happens

An open run refreshes every second: each server’s status, the step it has reached and the output it has printed so far. Only output that has not been shown yet is fetched, so a long log does not slow the page down.

A run waiting for someone to confirm its canary server says so at the top, with the canary’s output available to read before the rest of the servers start.

Runs are never edited

Runs are not changed after they finish. Retrying failed servers creates a new run linked to the original, so the first run still shows exactly what happened, and the retry shows what happened the second time.

Running a command again from its history fills in the same command, account and time limit, but not the servers. Those are chosen again on purpose, so a command is never repeated on a fleet by accident.

Search and filter runs

Filter runs by type (playbook, template or command), by status, by playbook or by server. Filtering by server shows everything that has been run on that machine.

Open a run to read any server’s output, or download a server’s log as a text file.

Accurate status for every server

A run is marked as succeeded only if it succeeded on every server it targeted. A run that worked on some servers and not others is marked as partially failed, with the servers grouped by outcome.

A server that was never changed says why: blocked by the preflight, skipped after an earlier failure stopped further batches, or cancelled. The difference matters when you decide what to retry.

Private to your account

Everyone on your account can read every run, including commands and their output, which is why commands should never contain passwords or tokens. Nobody outside your account can open one of your runs, even with its address.

Frequently asked questions

Are dry runs recorded?
Yes. A dry run is recorded like any other run, marked as a dry run, with what it found on each server.
Can I see everything that was run on one server?
Yes. Filter the run history by server, or open it from that server’s page.
Can a run be edited or deleted?
No. A retry creates a new, linked run, and the original stays as it was.
Can I download the output?
Yes. Each server’s log in a run can be downloaded as a plain text file.
Does the history show who ran something?
Yes. Every run records the person on your account who started it, and confirmations are recorded the same way.

Key points

  • Who, what, parameters and servers for every run
  • Per-server checks, steps, output, exit code and duration
  • Retries create linked runs; nothing is overwritten
  • Filter by type, status, playbook or server
  • Download any server’s log

Related features