Skip to main content

Schedule history and reports

Every run leaves a report. This is where you read them, trigger one on demand, and turn a finding into a card.

Where to find it: a schedule's full history at Workspace → Operation → Costs, then click a schedule card's View full history → (/operation/costs/:cronJobId). An individual run's report lives at /operation/costs/:cronJobId/runs/:runId.

The full history screen​

A ← Costs link at the top returns to the main Costs page.

The header shows the schedule's title, a Paused badge if it is disabled, its recurrence in plain language, and either Next: <date/time> (enabled) or "Next: Paused" (disabled).

If the schedule cannot be found (for example, it was deleted), the page instead shows "Schedule not found." with a Back to Costs link.

If you can manage checks, two buttons appear next to the header:

  • Pause schedule / Enable schedule: toggles the schedule on or off.
  • Run now: triggers an immediate run outside the regular schedule. While starting, it reads "Starting…" with a spinner. If triggering fails (no credits, past-due subscription, project budget limit, or a temporary outage), a red "Failed" banner with a specific message appears at the top and clears itself after 10 seconds (or dismiss it with its X).

Below that, a Post results to communication channels switch (on a card of its own): "Applies to scheduled runs only. Run now always stays inside ACE, and changing this does not close an existing thread." Disabled if you cannot manage checks.

If you can manage checks and the schedule has a prompt, an Agent prompt section appears as a collapsed disclosure. Click it to expand and read the exact prompt text that gets sent to the agent on every run.

Run history​

A Run history heading shows a count badge and, while the list is refreshing, a small spinner. Below it: "One card per execution — click for the full agent report."

If there are no runs yet: "No runs yet. Use Run now or wait for the schedule." Otherwise every run for this schedule appears as a card, newest first, and the list automatically refreshes every few seconds while any run is still in progress:

  • In progress: a spinning loader icon, the run's source ID (monospace), and "Running" in the primary color.
  • Failed: a red dot, the run's source ID, "Failed" in red, and the error message below if one was recorded.
  • Complete: a colored severity dot (red = critical, orange = high, yellow = medium, green = low), the source ID, and the run's type if set.

Every card also shows the run's title (or a fallback: "Running scheduled check…", "Scheduled check failed", or "Scheduled check"), its source, and how long ago it ran ("In progress…" while active). Click anywhere on the card to open its full report.

How do I open a run's report?​

A ← Costs link returns to the main Costs page. The report header shows the run's title (or "Running scheduled check…" or "Scheduled check" if it has none yet), when it ran, its type, and its source ID in monospace.

If the run cannot be found: "Run not found." with a link back to the schedule's history.

Create Card / Linked issue​

Next to the header, if the run has not been turned into a card yet and you have permission (both the same manage permissions as above, plus issue:create, and the project is not read-only), a Create Card button appears.

Clicking it opens "Create card from cost report":

  • Title: pre-filled as [Cost] <run title> (editable).
  • Additional instructions (optional): a text area: "Add constraints or acceptance criteria. The schedule and full report are included automatically."
  • Cancel closes the dialog without creating anything. Create card is disabled until Title has text, shows a loading state while submitting, and closes the dialog on success. If creation fails, an error message appears above the buttons.

If the run is already linked to a card, the button is replaced by "Linked issue:" with a link to the card (its key and title, if you have permission to read issues), or just "Development card" as plain text if you do not have that permission.

Discuss in chat​

If a run exists, a Discuss in chat button appears (with a message-square icon). Clicking it opens a chat conversation pre-loaded with this run's context and navigates you to /chat. While opening, the button shows a loading state. If it cannot open the conversation, a red error message appears below: "Could not open this scheduled check conversation."

Report states​

Below the header card, the page shows one of the following:

  • Running: a large spinner, "Agent is running…", and "The full report will appear here when the check completes. This page refreshes automatically every few seconds."
  • Failed: a red-tinted box: "This run failed before a report was produced." plus the error message if one was recorded.
  • Complete, with a report: the agent's full write-up, rendered as Markdown.
  • Complete, no report: "No report was produced" and "This run completed without report content. You can discuss the execution in chat or return to the schedule."
On this page