@masek We have (had?) stuff in me-central-1. While one AZ was offline some redundancy for us got lost, but everything basically kept working as normal.
But once the second AZ went away we are basically dead in the water there - while some stuff still responds, basically nothing really works - the API+EKS controlplanes do not respond, elasticcache/redis complaining about being read-only etc.
Starting new instances in the surviving AZ is also no longer possible.
The majority of our affected services have been migrated to other locations so far, most back into our own infrastructure in Europe - those were mainly with AWS-UAE to be more local to some customers in the area.