Using Root Cause Analysis to Investigate Application Issues

As an IT manager, diagnosing a problem isn’t always as easy as it sounds. Imagine a typical scenario: a user calls you with a problem. Maybe they’re unable to access their email, or they’re getting an error code when trying to load a particular website. Unfortunately, the user can’t tell you what the problem is — they can only tell you symptoms of the problem. Sometimes these symptoms point to an easy fix, and other times you have to dig deep to discover what’s actually going wrong in your infrastructure.

Let’s consider an instance when a user calls in to report that the internet is down. This can be a result of anything from a downed server to a problem with virtualization. The sheer number of possibilities is endless, and as any IT manager knows, sifting through the possible problems can be both frustrating and time-consuming.

This is where root cause analysis (RCA) comes into play. This type of analysis involves the use of a specialized software that can quickly and efficiently determine the origin of an error or problem.

The key advantage of RCA is that it doesn’t just treat the symptoms — it identifies and resolves the underlying problems causing the symptoms. This can save you significant time, while also preventing costly downtime for your application or website.

The Domino Effect

RCA operates under the assumption that all actions and events are related. In other words, every action a user takes causes another action to occur. By going back and analyzing this string of events — similar to following a trail of breadcrumbs — this type of tool identifies the originating action that triggered the chain.

On a fundamental level, RCA follows five main steps:

• Identify the symptoms (such as slow or no internet, error messages, etc.).
• Collect data and facts by talking with the people involved, from end users to technology experts.
• Determine all possible causes of symptoms. This is achieved by examining the series of events and conditions surrounding the problem. RCA tools dig deep and dissect a problem to reveal its singular components. Charts, diagrams and other visual aids can be helpful here.
• Determine the one specific, main root cause.
• Pinpoint what action to take to prevent the symptoms from recurring. This step involves choosing a solution, evaluating the risks, implementing it and designating a point person to manage it.

General Causes of System Issues

As users of RCA tools trace back a series of events, they typically encounter these primary types of originating causes:

• Human error: This is where a person took an erroneous action, or took no action when one was needed. For instance, perhaps someone forgot to schedule a data backup or chose the wrong option on a software menu.
• Physical failure: This describes an instance when a system component failed or malfunctioned in some way. An example might be a server going down or data being breached. In many cases, human error leads to physical failure.
• Company-wide causes: This describes when a flawed or nonexistent organizational process leads to a system problem. For instance, if a company doesn’t have a data backup process in place, the lack of policy could lead to irretrievable data loss.

An RCA tool examines all of these origins, identifying any trends in user behavior or glaring system issues. In many cases, the analysis finds that more than one cause is responsible for the problem.

Example Solution: SolarWinds

SolarWinds has created a software package that can address these issues. This server monitoring software uses two protocols to determine the root cause of any problem: the Application Performance Monitor (APM) or the Synthetic End User Monitor (SEUM).

The Application Performance Monitor works from your end to double- and triple-check every aspect of your system. The APM monitors the performance of your application, as well as each individual component of the application, so you can get a bird’s-eye view of what’s gone wrong.

The Synthetic End User Monitor functions as a virtual user to get a deeper understanding of what your live user is actually experiencing. This can be particularly helpful with monitoring website performance, enabling you to diagnose the problem quickly and efficiently and avoid extended downtime. Several other companies offer this software, all with very different features; careful investigation is required.

No matter how simple or complex your system actually is, the key to a successful diagnosis is achieving a holistic view of every component of your application or website. Your infrastructure is unique, so it’s important to have a monitoring software that can intuitively navigate and adapt to your application environment.

Whether you’re frustrated with an overly complicated IT process or you want to proactively plan for the future of your application or website, a solution like SolarWinds enables you to cut back on time spent troubleshooting and focus instead on the solution.