Cluster 1 Hosting Update: Monday 25 November
The actions that we put in place on Friday appear to have maintained our services over the weekend.
This morning (Monday), one of the three sub-Clusters is seeing reasonably high CPU usage, currently peaking at just over 60%: this is 30% higher than the next busiest sub-Cluster, and 50% higher than the lowest.
There are only 5 installations on this busy sub-Cluster (A), and so we are now better able to focus on understanding where the root cause lies.
Until the busy period ends — usually around 11am — we will continue to monitor the Cluster, and take remedial action if required (by simply switching one or two of the sites to one of the less busy sub-Clusters).
Any further actions that we take will be agreed by the team, and actioned only after the morning traffic spike has levelled off.
We will continue to post updates here, as appropriate.