26 Jul 2016
16:00 UTC
Paul Cammish
This issue should now be resolved for all affected users.
The root cause of the problem was a single customer volume which was in effect consuming all available IO on the server, leading other volumes to have excessively slow performance.
This was unfortunately compounded by a rebuild of the physical disk pool due to a failed disk which had been swapped out earlier in the day.
Once the volume in question was identified and migrated away, load returned to normal, and we will be looking at adding monitoring to identify similar situations before they become service affecting.
26 Jul 2016
10:07 UTC
Bytemark Engineer
Some users on the cloud hosting platform may be experiencing slow IO problems due to temporary heavy load. Affected users will notice reduced IO performance on their servers. Engineers are working on resolving the issue currently.