Monitoring jobs
The Jobs page is where you monitor spinner runs for a project. Each job is one execution of a published spinner — often across many contacts.
This page is for supervisors and builders who need to know: is the run moving, how long is it taking, and which records failed?
Open Jobs
Section titled “Open Jobs”Inside a project, open Jobs in the project navigation.
src/assets/screenshots/08-jobs-page.pngLayout
Section titled “Layout”| Side | Shows |
|---|---|
| Left | Runs — spinner name, progress, status, duration |
| Right | Records for the selected run (paginated) |
Select a run on the left to load its records on the right. From a record you can open outcomes and, when needed, the pipeline journey.
Refresh (on purpose)
Section titled “Refresh (on purpose)”Jobs uses an explicit Refresh control instead of aggressive auto-refresh flicker.
Why:
- Supervisors can read the page without rows jumping
- Large orgs avoid constant network chatter
- You choose when to pull the latest status
How to use it: click Refresh after you start a run, while waiting on a slow provider, or when Live Ops reports stalled jobs.
Live Ops itself may poll more often for summary cards; the Jobs page stays calm until you refresh.
Durations and timing
Section titled “Durations and timing”Each job shows timing so you can judge health at a glance:
| Field | Meaning |
|---|---|
| Started | When the run began |
| Ended | When it finished (or empty / in progress) |
| Total time / duration | How long the run has taken so far, or took overall |
Use durations to spot:
- Runs that should have finished minutes ago
- Sudden slowdowns after an AI or channel change
- Whether a “stuck” job is actually still processing a large table
Record-level timing appears in the pipeline journey when you need per-step detail.
Statuses you will see
Section titled “Statuses you will see”Exact labels can vary, but the ideas are stable:
| Status idea | Meaning |
|---|---|
| Running / in progress | Runtime is still working |
| Completed | Run finished (check failed record counts) |
| Failed | Run-level failure — open records and journey |
| Waiting | Paused on an external or human step |
A completed job can still contain failed records. Always glance at progress counts, not only the job status badge.
How monitoring fits with Records
Section titled “How monitoring fits with Records”| Screen | Best for |
|---|---|
| Jobs | Run health, batch progress, duration |
| Records | Per-contact outcomes for ops |
| Pipeline journey | Step-level debug for one contact |
Day-to-day ops often live on Records. Jobs is the control tower for “is this campaign run healthy?”
Success looks like
Section titled “Success looks like”- You start a run, refresh Jobs, and see progress increase
- Long durations prompt investigation before customers notice
- Failed records are opened in journey and fixed in the spinner
- Runtime is confirmed running when nothing progresses
Common problems
Section titled “Common problems”| Problem | Likely cause | What to do |
|---|---|---|
| Job never moves | Runtime service down | Start runtime; see Troubleshooting |
| Numbers look stale | Forgot to refresh | Click Refresh |
| Job completed but contacts wrong | Outcome / metadata issue | Use Records + journey |
| PLAN_LIMIT on start | Monthly jobs/records full | When a limit is hit |
| Only some records fail | Step or data issue | Journey on a failed record |