Service availability is monitoring regionally. For distributed services - CDN and DNS - this means we monitor those services using multiple geographically dispersed clusters of monitoring nodes. For non-distributed services like compute, this means we separately provision and monitor resources in each provider region. If availability is grouped by region, only one record is displayed for each service and the availability metric presented is the mean of all regions, and the service down symbol will be shown if any region is unavailable. Otherwise, separate records are displayed for each combination of service and region.