The Challenge

The customer had very highly skilled teams in each of their key domains: network, windows, unix, database, production support, and incident teams, but each team had its unique toolsets for managing their domains. The existing tools being used, such as Dynatrace, Solarwinds, Zabbix, and LogRhythm, provided rich data sets about the applications, Windows and Unix servers, the network, and the firewalls.

When a problem occurred, usually after a new production release, the subsequent root cause analysis was lengthy because the individual teams struggled to efficiently assemble a cohesive “big picture” view of the problem.  Too many tickets were created on isolated issues, making it challenging to manage the problem and address the underlying root cause.  Connecting all these events between various segments of the system became overwhelmingly complex resulting in longer times to resolution and degradation of services.  It was not uncommon for some outages to last an entire day.

Customer Objectives

• Improve customer experience and satisfaction
• Achieve higher levels of productivity and efficiency in their teams
• Reduce Mean Time To Resolution (MTTR)
• Correlation of issues from the application level, down through infrastructure and the network

The Selector Analytics Solution: A Single Pane of Glass for End-to-End Visibility

The Selector Analytics solution achieved the customer’s goals by seamlessly uniting data from their 18 different tools into a “manager of managers” view for a full representation of critical data. Selector Analytics incorporated application performance monitoring events from Solarwinds and Dynatrace, infrastructure alerts from Solarwinds and Zabbix, log data from EKS, LogRhythm, and Splunk, and network performance monitoring data from Solarwinds.

Selector Analytics then took that a step further by using correlation to understand the root cause of a problem.  It wasn’t so much about correlating a pair of events, but to correlate an alert with a change of state.  For example, showing a change in state in the physical infrastructure (e.g. a rise in CPU on a database server) and an application query not getting a response.  This helps TracFone prioritize which tickets to focus on to resolve the problem quickly and efficiently.

Of course, Selector’s correlation engine also does a great job of alert and event deduplication. In one particular case, TracFone’s infrastructure and 18 different tools generated 7,000 alerts and events, which Selector Analytics reduced to 4.

TracFone is about to embark on the next exciting phase of its deployment, which is to integrate Selector with ServiceNow and to trigger automation to remediate problems with minimal user intervention.

The customer gained valuable insight into their IT environment and reduced the time it took to resolve critical issues resulting in a greater level of system performance, reliability, and uptime. In addition, Selector Analytics provided TracFone the ability to drill down the correlation path for a more granular view of their events for improved visibility into their applications and infrastructure allowing for faster incident resolution. Not only did Selector Analytics achieve the desired results for the customer, but the solution required little effort and minimal disruption, allowing TracFone to focus on its core business objectives.

The Results

By bringing together data from multiple silos and establishing a single source of truth for event correlation and root cause analysis, Selector Analytics helped TracFone reduce MTTR and improve visibility into their IT infrastructure and applications using data from disparate sources to quickly identify root causes. As a result, TracFone benefited from greater system performance, reliability, faster incident resolution times, and reduced costs.