MONITORING & RECURRING WORK
Keep it alive.
Recorded in the Specstream frontend · Example data
Check again. Without asking again.
Start a questionChoose the work.
Recheck existing matches, discover new ones, or research additional fields. Start from a collection or saved scope.
Choose the rhythm.
Set the cadence, first occurrence, time, timezone, and validity window. Pause when the question needs a rest.
Choose the boundary.
Set maximum spending per occurrence and per period. Account, project, and authority limits remain part of the decision to run.
Every occurrence has a story.
Explore continuityKnow what happened.
See completed, partial, skipped, failed, and canceled occurrences. A skipped run should have a reason, not disappear into the schedule.
Inspect what changed.
Open the resulting work and compare it with what was known before. Missing current evidence does not erase an earlier observation.
Keep each result.
An occurrence produces connected work and its own fixed delivery. The schedule keeps going; the record stays inspectable.
A few useful answers.
Is monitoring a separate product?
It is the same Specstream work on a cadence, with explicit recurring authorization and limits.
Can I pause and resume?
Yes. Pausing keeps the schedule definition and its history. Resume or revise the scope when you are ready.
What happens when a limit is reached?
The occurrence must fit within its applicable authorization and capacity boundaries. The history should show why work could not run; there are no automatic overages.