default-hero.jpg

IT OPS USE CASE :

Proactively Identify Issues and Performance Problems

Overview

COMPANY TYPE

img-alertCable and internet services provider

USE TYPE

img-it-opsIT Ops

KEY BENEFITS

  • Analyze data too complicated to set thresholds for
  • Reduce the noise of alerts
  • Rank alerts by severity of impact

DOWNLOAD PDF

"Our level 2 support specialists hit the ground running when we get an alert from Prelert. If there are a broad mixture of metrics in the Prelert Anomaly Alert, then we immediately know the scope and the probable cause of what’s happening."

~VP of Technical Operations

The Company

Servicing customers in over 40 states, this company is one of the largest cable and internet service providers in the United States. This ISP holds customer satisfaction as a top priority, as millions of customers across the country rely on their entertainment and communication services.



The Problem

The operations team at this ISP was in charge of monitoring the status codes of millions of cable set-top boxes across the country, but the sheer volume of possible status codes (ranging from a simple notification that the customer’s On Demand queue is 80% full to an error code signifying they were unable to purchase a PPV show), users, and geographical areas made it impossible to try to manually baseline the data. The team was tasked with putting thresholds on the data and quickly saw it wasn’t possible. They were bombarded with tens of thousands of alerts each day and struggling to identify and prioritize the alerts that represented issues which could impact a larger number of users and impact customer satisfaction. For example, if video On Demand failed in one geography, it would represent a potential huge loss in income both from customers unable to purchase shows or movies on demand, as well as from disgruntled customers overwhelming the customer support hotline. The bottom line: The ISP could not physically threshold their data in any realistic manner or sort through the clutter of status code alerts to identify alerts that represented real service issues.



The Solution

Since 2013 the operations team at the ISP has been using Prelert’s Anomaly Detective® to better understand status codes and errors that have a wider impact on their user base, reducing the noise of these alerts to more accurately identify problems and improve customer satisfaction. The ISP’s VP of Technical Operations was excited from their first trial of Anomaly Detective: “We knew that Prelert was on to something when we saw their ability to automatically plunk out the anomalies in our real-time data without any effort. But we were really excited by its ability to pull related anomalies out of multiple data types simultaneously.”



The Results

Anomaly Detective helped the team at the ISP reduce the noise of their status code alerts and prioritize and rank problems by severity, impact, and rarity so the teams knew where to focus their time. It even identified rare events and status codes that the support team didn’t even know existed.

The software allowed them to see:

  • Fewer, more accurate, and more important alerts:
    • The VP of Technical Operations explains, “Our level 2 support specialists hit the ground running when we get an alert from Prelert. If there are a broad mixture of metrics in the Prelert Anomaly Alert, then we immediately know the scope and the probable cause of what’s happening.”
    • The teams could reduce the time spent guessing at the problem and focus on the truly valuble and important tasks
  • Correlated anomalies across multiple sources:
    • The software was able to easily flag unusual behavior across multiple dimensions of data, making it easier to identify performance issues clustered by geography
  • Actionable troubleshooting data included in each alert:
    • Anomaly Detective reports other correlated anomalous events in each anomaly alert, allowing the team to quickly find the root cause of service problems without hours of troubleshooting
Try It For Yourself

See what Anomaly Detective can do to your mountain of alerts and notifications. Download it for free and you can have it set up and automatically monitoring your data in the background in real-time within minutes. No more thresholds. No more manual calculations. No more learning about problems from your customers.

Automate data analysis, eliminate manual effort, and reduce human error.

Download free trial