Fiduciary Company Shifts from Reactive to Proactive Hardware Monitoring through Machine Learning
Executive Summary
A financial sector company lacked proactive visibility into its infrastructure, and its capacity to react was limited because they had to wait for an alert when their hardware capacity was reaching its limits. Panorama Technologies assisted in implementing a solution that would allow the infrastructure administrators and managers to:
- Have total visibility into the performance and availability of their technological components: machines, switches.
- Apply a predictive or preventive model to future machine disk usage.
- Proactively alert to prevent disk space shortages in advance.
Implement the functionalities of the Machine Learning application, which uses artificial intelligence concepts.
Obtain the following visibility for each machine:
- Total disk capacity
- Behavior prediction at a future time (forecast)
- Behavior and average time elapsed
Initially, behavior data from a certain period in the past was used for each disk. This data is crucial for the artificial intelligence to apply data forecasting.
The “Predict Numeric Fields” functionality was used as it allows prediction experiments based on a codebase. Once the experiment was completed, the application helped generate the code.
- Implement Capacity Planning (Machine Learning) for the disk components of all devices and platforms where it can be applied.
- IT Infrastructure and Operations.
-
Windows (System Logs)
-
Linux (System Logs)
-
Appliance (Availability Scripts)
-
Service Availability (server scripts)
-
Capacity Planning (Machine Learning)
The implementation of Splunk’s Machine Learning allowed the consolidation of all information in a single point, enabling the future visualization of disk behavior for all components.
This allowed infrastructure administrators to proactively address potential incidents caused by disk space shortages.
