All times in this report are in BST.
Between around 10:00 and 17:00 customers faced intermittent issues accessing a number of our services, including diaries and widgets. Each time this happened, the affected sites became available again within a minute or two.
After investigation, we have discovered that this issue was caused by network issues at Azure that started at around 10:05 and lasted until around 15:00.
During the incident, we contacted Azure to try to get more information about what was going on. While we were waiting for a response, we decided to investigate possible mitigating measures we could take to reduce customer impact.
At around 13:51 we deployed changes to login.resdiary.com and to api.resdiary.com to try to mitigate the situation. After further monitoring, it appeared that the mitigation was working successfully, so we decided to roll it out further to the rest of our applications.
Unfortunately the measures we took may have caused some customers issues with accessing www.resdiary.com, Classic Widgets in Southeast Asia and North Central USA, and Diaries in East Asia. We sincerely apologise for these problems, but at the time we had to make a decision since we were unsure how long it would take Azure to solve their network problems.
We will investigate whether we can avoid customer impact in future when taking this kind of mitigating action.
Further information about the Azure issue can be found at https://azure.microsoft.com/en-gb/status/history/ under "Network Connectivity - Increased Latency", dated 13/5.