Improved OCI Stack Monitoring now supports Baseline and Anomaly detection for additional metrics. That means performance problems for out-of-the-box resource types and new resources imported into Stack Monitoring can be identified sooner. The new anomaly detection capabilities include Prometheus-based resources and OCI Services (such as Load Balancer, Block Storage, MySQL, etc). These new capabilities add to the previous released OCI Stack Monitoring that supported anomaly detection for the out-of-the-box resource types (Oracle Database, WebLogic, Host, etc).
A baseline is a reference point for comparing the current performance of a resource with its previous performance and setting appropriate thresholds for performance metrics for the baseline. Based on the observation of performance metrics over time, baselines are calculated by applying machine learning algorithms.
In performance charts, anomalous metrics are visually highlighted if they are outside the normal range. Over time, the baselines will be fine-tuned as the system is used.
Enable Stack Monitoring Enterprise Edition to identify resource baselines
Baselines can be enabled on a resource by enabling Stack Monitoring Enterprise Edition in the licensing option. In the case of newly discovered resources, baselines will become effective after at least two hours from the time the resource is discovered, baselines become more accurate over time.
In addition to the default metrics with anomaly detection, you can select up to 5 additional metrics for each resource type. The additional metrics chosen are available to all resources of the same type within the same compartment. For example, baselines for IOPS on Oracle Database are enabled and all Enterprise Edition Oracle Databases in that compartment will receive the IOPS baseline enabled metric.
Baseline and anomaly detection for effective benchmarking
Baselining and anomaly detection are critical components in modern data analysis and system monitoring. By establishing a clear baseline, organizations can understand what constitutes normal behaviour for their systems and processes. This understanding allows for effective benchmarking, resource planning, and long-term trend analysis.
Anomaly detection complements baselining by identifying deviations from the established norm, enabling early detection of issues, enhancing security, improving operational efficiency, and ensuring quality control.
- Not yet a customer? Sign up for an Oracle Cloud Infrastructure trial account!