A member of a switch stack has crashed, which is causing some partial network disruption. We're currently investigating further and working to restore service as quickly as possible.
UDPATE 11:30: all affected traffic has been automatically routed via alternative paths and network performance should now be back to normal. We're continuing to investigate the cause of the failure and will provide further updates in due course.
UPDATE 1600: we have identified a power supply fault with the failed unit. The failed unit has now been removed and replaced with a new one and the stack is fully redundant again.