Amazon cloud to undergo global reboot
Amazon's entire global EC2 service - typically used to host websites and other Web services - is undergoing a rolling reboot that will take days to complete.
Rumours suggest the reboot is necessary in order to mitigate an underlying security bug in the Xen virtual platform upon which EC2 operates. Newer virtual instance types will not be affected.
Reboots affect different services differently of course. For mission-critical operations - say a Web or database server - you will want to pro-actively manage the process using load-balancers and/or multi-availability-zone master/slave configurations, so that specific services are not impacted as individual instances roll through their updates.
At Silverink, we use AWS services where cloud represents best value for our customers. We have already updated all affected instances ahead of the reboot, using exactly the procedures outlined above. Managed correctly, you should not experience any downtime.
Rather than respresenting a problem, we would argue this whole reboot process highlights the beauty of cloud services. We can scale any web service, mitigate any problem, handle any level of demand and avoid interruption - all whilst minimizing cost.