With the rising adoption of digital, cloud-based systems among consumers and organisations, the demand for data centres has grown tremendously over the past few years. In 2015, the market size for data centre operators in Southeast Asia alone was estimated at $1.2bn (source). Operators are increasingly feeling the pressure to deliver data in larger amounts, faster pace and more flexibly than ever before. There is a need for smarter data centre management solutions.
Eliminate the False Alarms
False positive alerts occur when users are notified of an event but upon further checks, the alert is actually unintended. In many systems today, false positive alerts are usually transient, i.e. they are triggered by temporary events and will not cause a persistent error in the system. For example, false positive alerts from temporary spikes in CPU utilisation, service restarts or server reboots. They occur due to the use of traditional approaches in system monitoring, such as user-defined thresholds or other basic automatic monitoring solutions.
N10’s anomaly detection module is able to to eliminate as many false positive alerts as possible while retaining all real alerts. It is enabled via our proprietary AI engine that continuously learn system behaviours from data and adapt itself accordingly. The engine can dynamically adjust the lower and upper bounds for its forecast models as system performance data is being streamed into it. Internal simulations (simulated with six months of performance data) evaluate the detection accuracy of N10’s AI engine to be approximately 95%, with a high F1 score of approximately 0.9. The elimination of false positive alerts enables a cost-effective system management process and high service level for data centres.
View relevant case studies: