Activity and jobs
Anything that takes more than a moment is a job: installing an app, scanning a portal, taking a snapshot, deploying assets. Jobs run on a worker separate from the web server, so you can close the tab. Activity lists them all.
Job kinds
Section titled “Job kinds”| Job | Title on Activity | Can be cancelled | Emails |
|---|---|---|---|
| Install an app | Installing app in portal | No | On failure |
| Scan a portal | Scanning portal for assets | Yes | None |
| Take a snapshot | Taking snapshot name of portal | Yes | Start, finish, failure |
| Deploy assets | Deploying recipe to portal, or N assets from source to portal | Yes | Finish, failure |
Any job scheduled more than a minute ahead also emails when it starts.
Statuses
Section titled “Statuses”Queued, Running, Succeeded, Failed, Cancelled. The sidebar’s Activity entry counts queued and running jobs; the portal’s page shows its last five.
The Activity screen
Section titled “The Activity screen”Filter by search, portal, status, type and date range (last 7 days by default). Each row shows the job, its portal, status, when it started (or when it is scheduled), and how long it took. Watch opens a running job; Open opens a finished one.
A job’s page
Section titled “A job’s page”- Progress while running, with a label for the current step.
- Log: every step, with Show verbose for the detail and Copy. Install jobs stream the HubSpot CLI’s output here. Failed pushes include HubSpot’s reasons and the request that was refused.
- Results for a deploy: every item and its outcome, with links into the target portal.
- Cancel for jobs that allow it. A queued job stops at once; a running one stops at its next checkpoint. Install jobs cannot be stopped once running.
- Reschedule for queued jobs: pick a new time, or Keep current time.
Scheduling
Section titled “Scheduling”The snapshot wizard’s Schedule for later sets a run time. The time can be changed with Reschedule while the job is still queued.
Reliability
Section titled “Reliability”- Each running job holds a lease and renews it regularly. If a worker dies, the lease expires and the job is handed to another worker, or failed if it has no attempts left (“The worker running this job stopped responding.”).
- App installs get two attempts; everything else gets one.
- When the worker restarts, cancellable jobs are paused and resume shortly after.
- The same deploy, scan or snapshot cannot be queued twice while one is already queued or running.
Notifications
Section titled “Notifications”The bell in the sidebar collects job starts and results, app updates (“app 1.0.4 is available for portal”) and recipes that have changed since they were deployed. Mark all read clears the badge. Read state is kept per device.
Emails come from Trading Cards and link straight to the job.