Continuous and uninterrupted monitoring of infrastructure enables IT operations professionals to learn about each network component's needs. Over time with the growing know-how and expertise, an IT admin will have a deeper understanding of what is the optimum level of performance to be maintained for devices in the infrastructure. At the optimum level of performance, a device is maximizing its potential, contributing to productivity, while also not exceeding the recommended performance setting. Below the optimum, a device is not contributing its potential to the infrastructure, while above the optimum, a device is risking its health, longevity and eventual performance deterioration. Maintaining this fine balance is one of the challenges for IT admins.
In traditional infrastructure management, IT admins set thresholds to critical performance metrics on network devices. Setting thresholds facilitates seamless allocation of resources, to ensure peak performance.
However, manually configuring thresholds is a huge task in a large infrastructure. An IT admin can't be expected to know every minute threshold setting for each device in the infrastructure, it is not humanely possible. Manual threshold configuration is a challenge also because:
Thanks to OpManager Plus' adaptive thresholds feature, network admins can seamlessly perform the critical task of setting up threshold levels on infrastructure components. The adaptive thresholds feature is made possible through machine learning, which enables quick adaptation to changing performance metrics of network devices. By using advanced predictive algorithms and percentage-based calculations, it sets highly reliable values for metrics and uses them to set thresholds for performance monitors, making the task easier than ever before.
Traditionally, threshold configuration is determined by IT admins after arriving at a nominal value for a particular monitor of a device, based on previous trends and utilization patterns. With this information, three levels of network monitoring alerts are set up - Attention, Critical and Trouble. The three levels of alerts are assigned baseline values, which when exceeded, triggers the relevant alert. Threshold configuration changes are done at the device level and also done in bulk, applying to multiple devices.
The arduous task of threshold configuration takes a turn for the better, with OpManager Plus' adaptive thresholds feature. The large amount of historical trends and patterns that IT admins had to go through, now gives way to a much quicker, automated process by leveraging OpManager Plus' observability and advanced predictive algorithm. This algorithm reads performance statistics over time intervals, multiple network usage patterns and determines a "forecast" value for each monitor. The predictive algorithms will require about 14 days of performance data to provide forecast values. After the data is established and forecast values are set, OpManager Plus begins threshold-based monitoring, using baseline threshold values to control the frequency and criteria for the alerts being raised.
Once Adaptive Thresholds are turned on, setting up alerts is simple. Users only need to specify deviation values for each level of alert. If a monitor value goes above the configured deviation value for a particular level, an alert with the corresponding level for that monitor is raised.
Let's say for instance, the forecast monitor value for network bandwidth usage monitor is 50 and the deviation values for Attention/Trouble/Critical are set to 5/10/15. In this scenario, if the bandwidth usage value exceeds 55 (50 + 5), an alert with the criticality "Attention" is generated. Similarly, if the bandwidth usage value exceeds 60 and 65, alerts with the criticality "Trouble" and "Critical," respectively, are generated.
Note: The example mentioned above, is when the deviation is set in values. The deviation can also be configured using percentage. To learn more about how the threshold levels are calculated in terms of percentages, please visit this page.
The net benefits of Adaptive Thresholds are:
OpManager Plus' Adaptive Thresholds feature, enabled by machine learning, empowers network administrators to seamlessly establish threshold levels for infrastructure components. Through the implementation of cutting-edge predictive algorithms and percentage-based calculations, Adaptive Thresholds offer accurate values for performance monitors. This feature proves to be a transformative enhancement as it simplifies network monitoring by dynamically adjusting to changing performance metrics, enhancing observability across the network. Learn more about OpManager Plus