On August 30, 2023 at 08:30 UTC, Microsoft Azure experienced fluctuating power diminutions that caused a subset of cooling units in a data center in its Australia East region to go offline for nearly a day. To prevent hardware damage, selected compute and storage units were powered down. This impacted multiple downstream services at different levels of severity for varying amounts of time.
The impact of power fluctuations at source following a lightning strike impacted various services in Australia East, with the exception of Azure Kubernetes Service (AKS), which was affected in Australia Southeast as well, since it was dependent upon Australia East.
The image shows the level and duration of the event’s impact on various services. Severity is indicated by color, and reveals both Critical and Warning units. The latter includes, for example, disruption to autoscaling, which was reduced to an approximate 65% success rate.
.png)
For more information on the persistence and impact of cloud outages on businesses today, view Parametrix annual Cloud Outage Risk Report
.svg)
