Concanon Presents:

Observability Series
Part 2:
Application Performance Monitoring

tony-kustwan

By Tony Kustwan

In Part 1 of our Observability series, we looked at what Observability is and how it differs from traditional monitoring. We also explored the different telemetry types needed to obtain true Observability and the various tools to address the first aspect, Synthetic Transactions. In this post, the second part of this series will focus on learning more about the tools and capabilities available to address Application Performance Monitoring.

What is Application Performance Monitoring?

“The umbrella term used to describe a wide range of strategies and practices of continuously observing and tracking application performance and analyzing data to assess business impact”

Gartner defines Application Performance Monitoring (APM) as a monitoring software suite comprising digital experience monitoring (DEM), application discovery, tracing and diagnostics, and purpose-built artificial intelligence for IT operations. At Concanon, we believe that APM is an umbrella term used to describe a wide range of strategies and practices of continuously observing and tracking application performance and analyzing data to assess the impact on the business.

Here is a list of some top Application Performance Solutions:

Cribl AppScope

The latest new guy on the APM/observability block, developed an open-source product called AppScope, which provides application-centric instrumentation and data collection. AppScope brings one instrumentation approach for all runtimes; AppScope offers ubiquitous, unified instrumentation of an unmodified Linux executable for single-user troubleshooting or distributed deployments. No language-specific agents or bindings are required.

Pros

  1. It is free
  2. Easy to use

Cons

  1. Only available for *NIX systems
  2. No fundamental reporting capabilities outside of the data collection

New Relic

A software analytics and application performance management solution for Telemetry Data collection, exploration, and alerts on all your metrics, events, logs, and traces from any source with our unified telemetry database.

Pros

  1. Insights is an excellent product and has many capabilities
  2. Generous amounts of plugins to extend data collection
  3. Excellent training and documentation materials are available
  4. Good at monitoring URLs, with specific criteria

Cons

  1. Requires a significant number of manual configurations to implement and support
  2. Product requires a significant amount of time to achieve value
  3. No alternative solution to consolidate alerts across implementation as opposed to making a significant investment in Insights to manually achieve this
  4. Lack of data fidelity

DataDog

A monitoring service for cloud-scale applications, monitoring servers, databases, tools, and services. It encompasses infrastructure monitoring, application performance monitoring, log management, and user-experience monitoring. DataDog is best known for its ability to unify disparate data sources, rapidly resolve performance issues, and detect threats in real-time.

Pros

  1. Datadog is highly effective at monitoring infrastructure
  2. This product has a vast amount of integration points
  3. Integrated Runbooks give guidance on how to respond to alerts

Cons

  1. Lack of Real User Monitoring (RUM)
  2. Lack of major APM functionality
  3. Limited depth in correlation in-app & microservice monitoring

Dynatrace

DynaTrace is an Application Performance Monitoring tool (APM) meant for collecting application performance metrics and extending their capability to monitor at an instance level. Dynatrace believes that its product is a single APM solution that delivers full-stack observability and that it can monitor all types of observabilities use cases quickly and efficiently.

Pros

  1. Synthetic checks and real-user monitoring
  2. Leader for APM solutions for 10 consecutive years
  3. AutoDiscovery, vMotion events detections, self-healing at scale
  4. Support of old technologies

Cons

  1. ActiveGates plugin is not as robust as the OneAgent
  2. Lack of broader reporting capabilities
  3. Limits the number of traces ingested per process, leading to missed anomalies and limiting the scale
  4. Lack of data fidelity

SignalFX

A real-time observability solution for cloud-native technologies such as microservices, serverless functions, and container orchestrated environments like Docker and Kubernetes. SignalFx gives IT organizations a platform that allows them to observe data in real-time, no matter the data source or type of application.

Pros

  1. Real-time architecture reduces Mean Time to Detect enabling DevOps to respond within seconds
  2. NoSample Distributed Tracing provides full-fidelity visibility
  3. Instrumentation is lightweight, based on open standards.
  4. Provides intelligent alerting, incident response, post-mortem, and collaboration capabilities

Cons

  1. Still considered a newcomer to this APM space.
  2. Only a SaaS product
  3. Without Splunk Core, you lack data fidelity
  4. Currently lacks Real User Monitoring

AppDynamics

A leading Application Performance Management (APM) solution. It is a tool that monitors your application infrastructure and gives you code-level visibility for highly distributed applications through transaction flow monitoring and deep diagnostics. AppDynamics believes its product offers seamless traceability and a view that effectively bridges both the APM and the Business product usage.

Pros

  1. Code Level visibility
  2. Synthetic and Real-User Monitoring
  3. Session Replays

Cons

  1. Lack of visibility into Virtual Environments and Network monitoring
  2. Alerts tend to generate false positives often
  3. AppDynamics hosting solutions do not provide adequate notice when the service is down

Why is Application Performance Monitoring not enough?

A performance monitoring tool is a must-have component of any Observability solution your organization implements. APM’s are great at their specific tasks; however, the idea that a solution can auto-detect/discover context about how an enterprise-class system is interconnected and determine how critical each piece is in relation to others is flawed. An AI tool cannot determine what is valuable to a business. It can only detect what it perceives as normal behavior versus what is outside of that norm. However, it cannot tell how valuable that particular information is to the health of the service you’re monitoring or how it interacts with other tool sets.

To Conclude

In the second post in this series, we looked at what Application Performance Monitoring is and the different application performance monitoring solutions. We also looked at why application performance is not enough, and finally, we looked at some of the pros and cons of these solutions. What’s next? In part 3 of this series, we will explore log monitoring, the third type of telemetry needed to achieve end-to-end observability.

Need help gaining Insight in your Apps?

Concanon provides digital transformation and observability consulting to help you gain confidence and trust in your analytics. This is through applying the best of breed tooling to gather the right data for the right problems. Reach out today to find out more.

Concanon's Author

Tony Kustwan is the Director of IT Operations at Concanon LLC, a big data solutions company. He has over 20 years of experience managing and leading global IT initiatives in some of the largest environments.  He thrives on solving a customer’s greatest challenges with simple and repeatable frameworks to collect, process, and enrich data for decision makers.