UPDATE: this is now resolved. Changes have been implemented so that the issue does not happen again.
I/O slowdown has been detected in one of our data storage clusters. We are working on resolving this issue.