Following is the Pingdom monthly report for a single instance of the OakLeaf Systems Windows Azure Table Test Harness running in Microsoft’s South Central US (San Antonio) data center:
The July 2010 downtime data is similar to that for preceding three months of Azure’s commercial operation, which began on January 4, 2010:
|2010||Jan||99.75%||1h 50m 02s||13||735ms|
|2010||Feb||99.99%||0h 04m 19s||1||734ms|
|2010||Mar||99.98%||0h 10m 00s||2||737ms|
|2010||Apr||99.71%||2h 05m 01s||6||782ms|
|2010||May||99.65%||2h 29m 59s||8||669ms|
|2010||Jun||99.65%||2h 29m 59s||8||660ms|
|2010||Jul||99.70%||2h 15m 01s||11||526ms|
Note that the OakLeaf application is not subject to Microsoft’s 99.9% availability Service Level Agreement (SLA), which requires two running instances.
It’s not known how many, if any, of the 11 outages were planned for, as an example, fabric operating system upgrades. Lack of system upgrades during February and March would explain the higher uptime percentage.
It will be interesting to see if the 28.5% reduction in response time from the first quarter to July continues into autumn.