Choosing the best application performance monitoring tools for your organization: A primer

May 22 2019
 

Juggling too many APM tools and too few actionable insights? Chances are, you can whittle your dozens of monitoring products down to just three. Here’s what you need to consider first.


How To Choose Application Performance Monitoring Tools

Application performance monitoring (APM) tools have become essential to enterprises as their technologists deal with increasingly complex apps. Of course, as you might have experienced, dealing with a complex set of disparate tools across siloed teams doesn’t make things any easier.

The answer: a one-stop solution containing everything you need — for complete visibility — now and in years to come. But what functionality do you need among the many new advancements on the market today?

In this post, we’ll walk you what to consider to build a toolset that works for your business. Let’s first step back and clarify the difference between APM tools and solutions.

What are application performance monitoring tools?

APM tools are the components that comprise complete application performance management solutions. Each tool works in tandem to deliver the various insights businesses need to ensure the health and availability of their software. APM solutions differentiate themselves with unique sets of monitoring tools, but the best include three in particular: business-impact analytics, end-user monitoring, and artificial intelligence.

We’ll get into those shortly. But before diving into a full comparison of APM tools, you have a few baseline criteria to assess.

10 key considerations

Starting with a handful of potential solutions, check which meet your needs in terms of:

1) Programming language support

What types of technologies do you need to monitor? .NET? Ruby? PHP? Do your developers need a solution that provides the flexibility to add more languages?

2) Cloud support

Which cloud provider best serves the cloud-based services that you use? Here’s a quick overview of the most popular:

 

Pros Cons
Amazon Web Services (AWS) The go-to provider for most enterprises (as the oldest and most mature)
Discounted bundle pricing packages
Confusing cost structure if you’re not sure how much infrastructure you need
Deployment requires experienced IT personnel
Heroku Beginner-friendly and easy to set up
Supports autoscaling
Expensive and slow for large or high-traffic apps
Limited in certain instances
Azure A close, enterprise-ready competitor to AWS
Comes with out-of-the-box integration with Microsoft products
Customers have cited issues with technical support and documentation
Lacks full breadth of IaaS and PaaS service offerings
Google Cloud Platform Intuitive, easy-to-understand pricing and UI
Industry-leading AI tools
No enterprise focus
Offers fewer features than its counterparts
Pivotal Cloud Foundry Privately contained cloud platform
Run your own cloud platform and don’t have to start from scratch or migrate workloads
Still need an IaaS to trade the capital expense model for the variable expense model
Alternate IaaS platform allows for economies of scale

3) SaaS vs on-premise deployment

Do you need SaaS-only, on-premise, or options for both? (On-premise might make sense if you have a separate intranet to monitor, for example, or SaaS when on-premise resources are limited.)

4) Ease of use

An APM solution that delivers value from day one is an absolute must. For all internal users (more on stakeholder requirements later), the solution should be fast and easy to:

  • Deploy: Can you get up and running quickly without installing anything onto your servers?
  • Use: If out-of-the-box, do the dashboards provide intuitive, easy-to-understand insights at a glance? Or do you need custom dashboards?
  • Maintain: How easily can you automate routine activities? How often does the system need to be patched or upgraded?

 

5) Reporting and analytics

Does the solution measure the KPIs you’re interested in? For example, do you plan to go deep into code-level diagnostics or track performance by certain parameters (like user location)? Most importantly, does the solution roll these insights into meaningful (yet simple!) reports?

6) Architecture

Is the architecture robust enough to scale easily? Is it flexible? Or does it pin you down to proprietary hardware or handle a limited number of metrics or applications?

7) Agent vs. agentless

Agentless monitoring (a bit of a misnomer that refers to agents embedded in the software) requires no separate installation or licensing. But perhaps you need proprietary agents for more intelligent functionality?

8) Security

Do you rely on third-party hardware or software? If so, they can weaken application security. What kinds of security features does the solution include? Permissions? Encryption? Periodical updates?

9) Pricing

Is the pricing flexible (per month per server vs site license)? What’s included and what extensions might cost extra? Can you save with bulk orders or annual discounts?

10) Technical support

What support options are available? Are they 24/7? Is the documentation high-quality and readily available? Does the vendor position itself as more of a partner than a service provider to ensure the success of your product?

These 10 criteria should narrow down your search. But what really sets solutions apart, as we mentioned earlier, is their unique set of functionalities or “tools.”

How to choose APM tools

The first step to identifying the right tools is to identify stakeholders and their critical pains.

In other words:

  • Who will benefit from APM? Look both within and beyond IT, from management to the end user.
  • What problems do they need to solve? Look at the rate and types of problems, their severity, and the number of employees involved in troubleshooting them.

This can help uncover blind spots in your existing environment. Added to the service level standards you’re legally required to maintain, you should end up with a list of requirements from which to identify must-have tools.

And we’d argue that a comprehensive APM strategy needs only three:

  1. Business-impact analytics to tie application issues to areas for business improvement
  2. End-user monitoring to identify issues before your customer does
  3. AI capabilities to address potential issues before they even occur

So what do these APM tools look like, and how can they impact IT?

Business-impact analytics

Some people call this “full stack tracing”: connecting the dots between app performance, user experience, and business outcomes by collecting, isolating, and correlating code-level details for transactions that take place in your app. The ideal tool automates this for every transaction and delivers insights in real time.

This goes beyond simply monitoring application performance to find out how that performance impacts business operations. Could it be a performance problem that caused a customer to abandon their shopping cart, for example? With that kind of insight, IT could inform the marketing team’s win-back campaigns and help drive superior customer experiences.

End-user monitoring

This ultimate measure involves tracking the actual experience of your users. What better way to find performance bottlenecks?

There are two main ways to do so: by simulating and testing synthetic user interactions and by passively monitoring the end-user experience as users interact with your applications in real time. The ideal solution offers both.

With these capabilities, app support teams can optimize every touchpoint in the user’s journey by capturing and fixing errors, crashes, page load details, and other critical metrics before the end user notices a problem. This can launch you miles ahead of the majority of enterprises that are alerted to anomalies by users.

AI and machine learning

AI-driven analytics are a relatively new frontier in APM, and has led to the advent of AIOps platforms. In recent years, they’ve developed to include machine learning, statistical analysis, and pattern recognition to automatically diagnose root cause and predict anomalous behaviors. For ease of use, the best APM tools distill these complicated insights into actionable information in a single pane of glass.

New though it may be, AI is indispensable in environments where apps are elastic, highly virtualized, and built with microservices. It can (and should) be a key differentiator in your search. By predicting events, you can squash possible issues before they become issues. Capacity planning becomes much less daunting.

Final thoughts

APM promises tremendous returns in terms of reducing MTTR and boosting the credibility of your business — unless you’re struggling to make sense of data from dozens of products across teams. Keep things simple by sticking to solutions that offer the core capabilities above.

Once you’ve honed in on two or three vendors and implemented a proof of concept, you’re ready to evaluate the strengths and weaknesses of their toolsets before picking your champion. And don’t just take our word for it! Always compare APM tools by putting them to the test.

We’re here to help if you need it, so come to us with any questions you have along the way. Good luck!

An introduction to AppDynamics

Adrienne Jack
Adrienne is a B2B content marketer who writes for tech and SaaS businesses that care about accelerating their customers' success. A few of her favorite things about sharing ideas online: learning something new every day, nerding out on data, and connecting with people all over the world. Say hello on Twitter @adriennegjack or learn more at adriennegracejack.com.

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form