Monitor
How to use Monitor to understand jobs, runs, refresh state, and operational failures.
What Monitor shows
Monitor is the operational view for recurring Stackless work. It helps you see whether scheduled jobs, managed model refreshes, dashboard exports, and other runtime tasks are healthy.
[picture 1. Monitor page showing jobs, recent runs, statuses, and timestamps]
Common things to check
Use Monitor when:
- A dashboard looks stale.
- A Transformation Model publish or refresh is still running.
- A scheduled export did not arrive.
- A source or model run failed.
- You need to know the last successful run time.
- The Agent says an operation was queued and you need its status.
Jobs and runs
A job is the recurring or queued unit of work. A run is one execution of that job.
Important run fields include:
- Status.
- Start time.
- Duration.
- Result or error.
- Related model, dashboard, source, or export.
- Whether a retry or manual follow-up is needed.
Reading statuses
Typical status meanings:
| Status | Meaning |
|---|---|
| Queued | Waiting for execution capacity. |
| Running | Currently executing. |
| Succeeded | Completed without a reported failure. |
| Failed | Stopped with an error. |
| Cancelled | Stopped before completion. |
| Paused | Not scheduled to run until resumed. |
Specific environments may display additional status details.
Troubleshooting stale dashboards
When a dashboard looks stale, check in this order:
- Source sync status.
- Transformation Model materialization and refresh status.
- Semantic Model publish state.
- Dashboard refresh state.
- Scheduled export run history, if the issue is an export.
The Catalog can help trace which upstream models and sources feed the dashboard.
Agent run behavior
Some Agent operations continue in the background after they are queued. If you navigate away and return to Agent, Stackless can reconnect to the active run or show the latest run state.
If a long operation appears incomplete, check:
- Whether the Agent is still streaming.
- Whether an operation ID or run appears in Monitor.
- Whether a private draft or artifact was already created.
- Whether the operation needs explicit confirmation before publish or refresh.
Escalating a failure
When reporting a failure, include:
- Workspace and environment.
- Dashboard, model, source, or artifact name.
- Run or operation ID, if visible.
- The approximate time of the failure.
- The user-visible symptom.
- Whether the issue affects one user, one role, or all users.
This information lets Stackless trace from the user-facing failure to the correct job, model, source, or Agent run.