Service disruption Jun 24
The service was entirely unavailable 2015-06-24 13:36 - 2015-06-24 14:30 GMT downtime 54 min. Before this there were failures in some accounts.
What was it
The Round Robin can do several attempts to assign a ticket to an agent in order to ensure assignment if an error occurs. If the assignment is not made within a certain time the Round Robin repeats the attempt. With increasing time of assignment over certain value, we got the situation in which the Round Robin tries to assign the ticket, while it has been successfully assigned.
After we had received complaints about incorrect app operation from some of our customers. We had turned off the ticket processing to prevent further impact on the customer's operation.
To fix the problem it took us 54 min, during which time the service was entirely unavailable.
What will follow
We will make an unscheduled release which changes which exclude the possibility of such re-assignment.