New software tools can help CIOs identify, fix, and prevent service-impacting glitches in real time.
- An observability solution enables IT teams to keep eyes on increasingly distributed and complex cloud environments and apps
- The software mines cloud data to pinpoint causes of customer problems that can otherwise go undetected
- Observability produces new, actionable insights about customer experience
As marketing director of technology for Stanley Black & Decker, it’s no surprise that Colleen Romero is into power tools. But her favorite utility isn’t a jigsaw or cordless drill.
It’s observability technology.
Observability—software that integrates data from various departments and applications —makes it possible to monitor how customers experience the performance of the 350 websites that Romero oversees. On her watch, the tools have spotted some problems before customers notice.
“We no longer have customers calling us to say our websites are down,” says Romero. “Observability tools let us spot any degradation of performance and identify the root cause so we can address it proactively.”
Observability enables businesses and IT teams to trace customer issues back to an offending application, process, or device, so that they can quickly address the issue and make changes to prevent it from recurring.
Before the emergence of observability software, most IT shops used a patchwork of point solutions to track metrics, logs and traces for individual departments. But the fragmented products were expensive to support and didn’t capture bigger-picture insights.
The latest observability tools can search many types of data—such as operational data logs and records from application performance monitoring (APM) systems—to spot relationships that no one had thought to track. A cable company, for example, might learn that increased customer churn was associated not with the usual suspects such as router problems, but perhaps to a glitchy upgrade of remote control software.
Observability drives better customer experiences
Observability can be especially critical in improving customer experience. If a website doesn’t work as advertised, 42% of visitors will leave and never come back; a site that takes too long to load can lose two-thirds of its visitors.
“You need observability tools in this age because the web is like a public utility,” says Romero. “It’s no longer OK for a page to load slowly or not at all. Customer expectations are far too high for that.”
Observability can also make IT operations efficient by directing staffers to the precise cause of a problem instead of forcing them to go on fishing expeditions using data that may not result in the right solution. Through additional data analysis, observability software also generates actionable insights for understanding the customer experience from the enormous amount of data produced by monitoring systems, says Jayne Groll, CEO of the DevOps Institute.
“You could easily get buried under mountains of data without observability tools,” she says.
For example, Furuno, a Japanese maker of marine electronics and satellite communications systems, recently used observability tools from Elastic to resolve complaints from shipping customers about spotty internet service and increasing monthly data charges. Previously these problems had been hard to troubleshoot because it took on average two days to get performance logs from satellite service providers, and even then the data came in the form of massive spreadsheets. It wasn’t obvious whether the problems stemmed from a broken antenna, bad weather, or sailors using excessive bandwidth to stream movies.
With observability, Furuno can now see an integrated picture of the various data sources, delivered to computers on customers’ vessels, which can lead to diagnosing why the internet was not working properly within an hour.
The observable future
Observability technology is positioned to make significant advances as ML capabilities are added. With current tools, it’s still up to human analysts to figure out how to respond to the tool’s insights, says Scott Sinclair, a senior industry analyst with Enterprise Strategy Group. Over the next few years, he says, machine learning capabilities will help observability tools diagnose failures and present solutions in real time, giving customer and IT teams a better chance to achieve a core objective—to keep more customers happier, around the clock, at times before they’re even aware of the problem.
“We’ve only scratched the surface,” says Sinclair.