<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=1110556&amp;fmt=gif">
April 8, 2022

The Future of Monitoring: Turning Unknown Unknowns into Known Knowns

by: Nishant Kabra

To ascertain risk, national security and intelligence professionals have long used concepts such as known knowns, known unknowns, and unknown unknowns. The idea of unknown unknowns was created in 1955 by American psychologists Joseph Luft (1916–2014) and Harrington Ingham (1916–1995). This concept continues to be used today in risk assessments and is applicable to technology. 

The unknown unknowns are the threats and potential problems that remain invisible until their impact manifests. When they do, they create consequences that can disrupt the status quo. 

In business, and in particular in technology, the unknown unknowns pose a serious problem and often remain elusive to IT teams. As applications have become more complex, it’s become harder to track risks. Traditional application performance monitoring (APM) has focused on monitoring for known problems. You knew the performance baseline for an application, as well as expectations and performance indicators, and you’d set thresholds and static alerts to see if performance deviates. You were monitoring for known knowns. Today, this is no longer sufficient.  

APM Needs to Address the Unknown Unknowns

Modern apps are extremely complex and multi-layered. You can no longer depend on monitoring for known problems. In the cloud, the application stack is so complex that issues can happen anywhere and disrupt app performance. These app disruptions can have a negative impact on the user experience, customer/user satisfaction, and even customer retention and business revenue. 

But can an APM solution really monitor for unknown unknowns? 

Yes, it can and here’s how. To monitor all the layers of the environment, and understand these layers and how they interact, an APM solution needs a topology model of dynamic and constantly changing environments. 

Along with a topology model, a solution must have the intelligence needed to determine when a metric deviates from norms. A solution needs to be able to correlate events and anomalies and identify probable root causes, then automatically remediate the issue. An APM solution such as Broadcom’s DX APM SaaS does this with a rich topology based on our strengths in monitoring applications, infrastructures, and networks. We have a built-in, automated remediation engine that leverages the topology and helps customers automatically remediate issues when they occur. 

DX APM SaaS is capable of dynamically monitoring for unknown unknowns, enabling businesses to elevate their level of observability and create confidence that they won’t run into surprises. 

What does this look like? Here’s a scenario to illustrate. In this fictional scenario, we’ll demonstrate how John, an SRE/DevOps engineer, can use our DX APM SaaS solution to query across captured instrumentation data to examine, correlate, and derive insight or uncover unknown unknowns (observability) from this data.

1. John is notified of a performance problem with the ACME Business Service. It appears that several users are unable to log into the site. John logs into DX APM SaaS and opens up the experience view for service health and notices that the service is indeed “Red.” Next, he drills down into service health. 

ESD_FY2022_The-Future-of-Monitoring.figure_01

2. He sees that the activity service application is reporting a problem with several failed transactions. He sees that APM has already detected a problem and is reporting an issue with the authenticator service. He clicks on the problem to view the associated details in a problem analysis notebook view.

ESD_FY2022_The-Future-of-Monitoring.figure_02

3. In the notebook view, John views the problem details and notices that there are issues reported with several different technology tiers of the application stack.

ESD_FY2022_The-Future-of-Monitoring.figure_03

4. He notices that the system has identified the authenticator service as problematic and selects the service.

ESD_FY2022_The-Future-of-Monitoring.figure_04

5. He opens up the traces tab to take a look at the automatically collected transactions and sees several error traces. He opens the first trace and sees that the request from the first service has timed out with a 400 error.

ESD_FY2022_The-Future-of-Monitoring.figure_05

6. At the same time, he also observes that only some transactions are timing out while others are successful.ESD_FY2022_The-Future-of-Monitoring.figure_06

7. He wonders what could have changed with the application and brings up the timeline view to see change events. He notices a new version of the authenticator service was released to production last night. The new version was re-implemented as a microservice in a new technology stack. Furthermore, the specific team that developed the new service decided to instrument it with OpenTracing for their observability needs. Could there be an issue with the new service?

ESD_FY2022_The-Future-of-Monitoring.figure_07

8. He turns his attention to the alarms. The system is reporting an abnormally high CPU utilization on the new service. In addition, the system has also detected a change in the deployment configuration, with the number of replicas going down from six to one.  This is obviously not right: the new service was checked in with an incorrect deployment configuration. This error caused the service to be under-provisioned, resulting in performance issues. 

ESD_FY2022_The-Future-of-Monitoring.figure_08

As you can see, DX APM SaaS automatically detects issues via dynamic baselines and correlates these seemingly unrelated sets of events, both performance and change events, together in a unified topology. This helps John quickly diagnose the problem, down to the specific line of code. As a result, the solution converts unknown unknowns into known knowns. 

Learn more about how DX APM SaaS can help you address unknown unknowns.  

Explore More Posts

View All Blog Posts
May 20, 2022

How AppNeta Drives Business Value

Learn how to tie AppNeta monitoring to business value by reading the core value areas AppNeta can provide and the business challenges these address. Read Now
May 20, 2022

Top 5 Reasons for “Why AppNeta?”

Here's how, unlike its competitors, AppNeta helps you gain invaluable insight into the end-user experience. Read Now
May 19, 2022

Monitoring Azure and Your Entire Hybrid Infrastructure with DX UIM

Find out how DX UIM enables teams to do efficient, comprehensive monitoring of their Azure environments and their entire hybrid, multi-cloud ecosystem. Read Now
May 12, 2022

In Digital Transformation, Don’t Overlook the User Experience

AppNeta for Symantec Network Security delivers end-to-end performance visibility. Read Now
May 4, 2022

Expert Series: Large MSP Was First to Upgrade to DX UIM 20.4

Learn how a managed service provider leveraged their DX UIM 20.4 upgrade to create dashboards, group servers together, and develop reports faster. Read Now
April 13, 2022

NoSQL Database Monitoring with DX UIM

This blog offers an overview of NoSQL databases and details a few of the most popular out-of-the-box DX UIM probes that are available for these databases. Read Now
April 8, 2022

The Future of Monitoring: Turning Unknown Unknowns into Known Knowns

Traditional APM has focused on monitoring for known problems. Today, that isn’t sufficient. See how you can monitor for unknown unknowns. Read Now
March 25, 2022

Visibility Anywhere: Key Takeaways from the NetOps Virtual Summit

Find out about some of the key takeaways from the 2022 NetOps Summit, which was centered on the theme “visibility anywhere.” Read Now
March 24, 2022

Do you have your hybrid cloud strategy all figured out?

As your organization grows increasingly reliant on hybrid cloud environments, advanced, scalable monitoring is vital. See how DX UIM can help. Read Now
March 11, 2022

Application Discovery with DX Unified Infrastructure Management

DX Unified Infrastructure Management (DX UIM) offers an open approach to application infrastructure discovery. Use scripts to discover more about a device. Read Now