That sounds like a UI/filtering mismatch rather than your runs not actually existing. Most likely, the dashboard filter is matching the state name

Running

too literally, while some “active” runs you’re seeing elsewhere may actually be in related states or variants that don’t match that exact label. A few useful details: * In Prefect, state type and state name are different concepts. *

RUNNING

is the underlying type. *

Running

is one possible displayed name. * Some runs that look active may actually be in states like: *

Retrying

with type

RUNNING

*

Submitting

/

InfrastructurePending

with type

PENDING

*

Late

with type

SCHEDULED

So if the UI filter is effectively doing “name equals `Running`”, it can miss runs that are still operationally active. What I’d try: 1. Clear all filters and re-apply only the running-related one. 2. Check whether those “running” flows you see elsewhere are actually labeled: *

Running

*

Retrying

*

Pending

*

Submitting

*

InfrastructurePending

*

Late

3. Try a different runs view, if available, to see whether the issue is specific to one page. 4. Refresh the page / hard reload to rule out a stale UI state. If you want to confirm whether this is a UI bug, it would help to compare: * a run that appears “running” in another view * the exact state shown on that run’s details page * whether it disappears only when the

Running

filter is applied I didn’t find a GitHub discussion specifically describing your exact symptom, but the state filtering behavior in Prefect does have some nuance around name vs type, and that fits what you’re seeing. Relevant references: * Prefect states concept docs: States * Prefect repo issues: GitHub issues If you want, I can help you narrow this down by walking through: * which dashboard page you’re using * what exact state label those “running” flows show on the run details page * whether this is Prefect Cloud or self-hosted UI behavior across all views

I think I first noticed this yesterday