All the alerts are marked as resolved in the Incident detail after some time. The scenario is the following:

1. Click on the active/firing incident and check that all the alerts are firing
2. deselect the incident and wait around 5 minutes
3. click again on the incident detail and notice that all the alerts are resolved now (but in fact they are not)

The issue seems to be that there's a new query for the alert data, but the range time (here remains the same and then the delta calculated at https://github.com/openshift/monitoring-plugin/blob/main/web/src/components/Incidents/processAlerts.ts#L171 increases and everything is hapily resolved.