Monitoring Or Observability? It Doesn’t Matter If You’re Measuring The Wrong Things

Innovation Monitoring Or Observability? It Doesn’t Matter If You’re Measuring The Wrong Things Mehdi Daoudi Forbes Councils Member Forbes Technology Council COUNCIL POST Expertise from Forbes Councils members, operated under license. Opinions expressed are those of the author. | Membership (fee-based) Jun 24, 2022, 09:45am EDT | Share to Facebook Share to Twitter Share to Linkedin CEO and Co-Founder of Catchpoint , a digital experience intelligence company.

getty If you were to go back in time four years and do a Google search for “application performance monitoring,” you’d see a list of companies offering APM services. Now, come back to the present and do another search for “observability. ” You’ll see most of the same names.

So, did these companies overhaul their businesses to offer different kinds of tools? Not really. The evolution of language around monitoring owes more to changing fashion than changing technology. Somehow, in the last few years, the term “monitoring” has come to feel outdated.

“Observability” sounds newer and more scientific. Look under the hood, though, and you’ll find the actual approaches aren’t all that different. Monitoring, as traditionally defined, measures predefined performance indicators from system components.

Observability aims to be more holistic, analyzing outputs of a system to infer its health. So, for example, if you wanted to measure a bakery’s cake-baking performance, monitoring might track metrics like baking times and temperatures, and observability might examine the aroma of cakes coming out of the oven. Philosophically, then, there are genuine distinctions.

But here’s the thing: It doesn’t matter. There will always be a role for monitoring KPIs in a digital business. And it’s entirely reasonable to infer the health of web applications from system outputs too.

But whether you’re “monitoring” or “observing,” none of it matters if you’re not measuring the right things. So, what should you be looking at? The actual digital experience of your customers. MORE FOR YOU Google Issues Warning For 2 Billion Chrome Users Forget The MacBook Pro, Apple Has Bigger Plans Google Discounts Pixel 6, Nest & Pixel Buds In Limited-Time Sale Event Understanding Observability The technical definition of observability—inferring internal health from external outputs—encompasses three components: 1.

Metrics , or collecting performance indicators like disk memory and throughput from system components. 2. Logs , which record what happened, where and when across a digital infrastructure and are critical when analyzing a system to understand what went wrong.

3. Traces , which dispatches software agents across systems to isolate problems. These are all valuable tools for a digital business, and you should use them.

The problem: They’re all passive. Fundamentally, they’re about reacting to issues after they’ve already affected users. Wouldn’t you prefer to know you have a problem before it disrupts your customers? Avoiding Blind Spots Whatever a monitoring framework calls itself, its value depends on how accurately it reflects real customer experiences.

Returning to our bakery example, it might be helpful to track how much flour you’re using or the color of batter at different points in the process. At the end of the day, though, what matters is whether people like your cakes. If your monitoring doesn’t tell you that, it’s measuring the wrong things.

How do you know when you’re not getting the full story? Watch for pitfalls like: • Overreliance On Thresholds: Large digital applications can generate millions of alerts daily—far too many to investigate each one. So, instead, monitoring services set thresholds, such as surfacing an emergency alert if an API fails 10,000 times. That’s fine for part of your monitoring strategy, but do you really want to rely on it to track the health of your digital business? Effectively, you’re saying if “only” 9,999 customers have that issue, it’s no big deal.

Wouldn’t it be better to know that something abnormal is happening sooner? • Overreliance On Logs: Logs play a critical role in diagnosing problems. But if users can’t get to your site, they may not generate any logs for you to review, so you’d never know something went wrong. • Limited Geographic Resolution: If you’re serving a global customer base, you have to account for all the markets from which users visit your site over different regional and last-mile networks.

Even if a monitoring service claims to cover a given region, that doesn’t mean they do it with any granularity. For example, an observability service might claim to track performance in Asia by measuring from Hong Kong. But there are 4 billion people in Asia, and most of them don’t live in Hong Kong! These gaps affect real digital businesses every day.

Whether you call the tools they’re using “monitoring” or “observability,” the point is they’re broken. As an industry, we can’t just keep slapping new names on what we’re offering while continuing to deliver the same mediocre results. Achieving True Digital Experience Observability At the end of the day, your monitoring tools should all be instrumented toward answering a single question: Can customers access my site and easily complete transactions? Keep that goal front and center, and your monitoring strategy gets simpler.

Yes, keep monitoring (or observing!) the health of internal systems so you can react if something goes wrong. But make sure you also do the following: • Look beyond your own systems. Modern web applications have grown enormously complex.

Hundreds of distinct components might be involved in rendering a page—most coming from third-party infrastructures outside your control. Problems with any one of those third-party components could degrade the user experience. But if you can only see your own infrastructure, you’d never know.

• Monitor from the outside-in. Even if internal systems look good, that doesn’t mean users aren’t having issues. Similarly, just because things look fine in one location, you can’t assume they’re fine everywhere else.

You need to maintain 360-degree visibility—measuring digital experience from the customer’s point of view. Collect data from everywhere you have customers. Start at the user and move back through the network, all the way up to application containers.

• Be proactive. If you’re waiting for problems to surface on passive instruments, you’re probably already too late. Web application issues propagate at the speed of data, so even small problems can quickly snowball into massive outages.

Instead, pair passive user monitoring with proactive synthetic monitoring. By generating “digital mystery shoppers” in all the markets you operate, you can proactively probe and poke at your application just as real users would. It’s the only way to spot problems before they disrupt real customers and transactions.

No one said monitoring a modern digital application was easy. But get these things right, and you’ll find yourself observing what matters most: great customer experiences and a healthy digital business. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives.

Do I qualify? Follow me on Twitter or LinkedIn . Check out my website . Mehdi Daoudi Editorial Standards Print Reprints & Permissions.