- Jason Rincker: Principal UX Designer
One of these things is not like the others!
Outlier detection allows users to automatically identify a device or group of devices that is behaving abnormally compared to its peers. Outlier detection is very effective in the detection of devices/components that are operating in a degraded state but not bad enough to be detected by a threshold alert.
In Digital Operational Intelligence (DOI) for a service, application or group, users can compare the web servers, app servers, databases, VMs, hosts, connecting n/w segments with each other and find the outliers that are experiencing degraded behavior. To achieve this, DOI monitors multiple metrics across the different device or component types.
Here is an example of a connection imbalance in server-focused data from Unified Infrastructure Management (UIM). The load balancer was set to do round robin connection balancing, but since the VMs running Nginx were sized differently, some ended up queuing connections while others were sitting idle. Such detection is a big-time save and improves for the efficiency of the SREs.