Availability metrics are reported at an aggregate level across all tiers and error types.
Individual customer availability may vary depending on their workload, autoscaling settings and API features in use.