We have identified a provider-related networking issue. This one had an effect on multiple system components, including database clusters.
It looks like Hetzner has already solved this issue.
Nearly all components were affected by this (delivery, UI etc.) for a short period of time.