Sunday, December 2, 2018

Anomaly detection system POC

Got an opportunity to work on time series data and Anomaly detection application on Data center.

This system is capable of monitoring health of servers and provides remediation whenever server is unhealthy. Our Intelligent system uses Machine Learning to detect or forecast Anomalies and warn the Service providers to ensure timely action

· We have designed the Orchestrator in such a way that it monitors those specific services which are added to our database. When any service goes down it raises a ticket, then it will run corresponding remediation solution to resolve the issue. Finally it closes the ticket after it completes the resolution.

· ELK Stack is used to collect the data/stats from specific server for monitoring health, in turn we use machine learning models to predict the anomaly in the data captured for the specific duration.

· In the similar way our machine learning model also forecasts the future potential anomalies based on specified criteria and raise a warning ticket for timely action hence it is capable of preventing upcoming issues
Used Plotly for visualizing the plots.
Below are some of the important screenshots