Why not apply machine learning to machine management?
An average IT or cloud operations team had to deal with nearly 3,000 daily monitoring and management alerts in 2019, suggests a report from American software intelligence company Dynatrace.
This represents a 19 percent increase from 2018.
The company estimated that the time spent dealing with alerts costs large organizations an average of $1.5 million in overhead expenses each year and suggested that employing artificial intelligence to identify critical issues is the only way to handle the growing complexity of multi-cloud deployments.
Too much, too fast
Dynatrace offers application performance management (APM) and cloud infrastructure monitoring products, and is a major player in the emerging AIOps software category that hopes to bring the benefits of artificial intelligence and automation to IT infrastructure management.
The company has just published a report titled “Top challenges for CIOs on the road to the AI-driven autonomous cloud.”
It surveyed 800 CIOs and found that an average IT team spent around 15 percent of its time trying to sort through monitoring alerts to decide which issues required immediate attention, and which could be safely ignored. In an IT department that pays its staff $10.2 million per year, this would amount to $1.53m lost annually.
Seventy percent of respondents said their organization was struggling to cope with the number of alerts from monitoring and management tools, and 75 percent said most of the alerts they received were irrelevant. On average, just 26 percent of the alerts received each day required actioning.
At the same time, an average organization experienced 21 incidents that could have been prevented if alerts were seen or acted upon in time but weren’t.
The situation is clearly unsustainable, and Dynatrace suggests that CIOs are increasingly looking to AI and automation to solve the issue of information overload.
“The Dynatrace Software Intelligence Platform is a single platform with multiple modules, leveraging a common data model with a precise explainable AI engine at its core,” said Bernd Greifeneder, CTO and founder at Dynatrace.
“Unlike other solutions, which just serve up more data on glass, it’s this combination that enables Dynatrace to deliver the precise answers and contextual causation that organizations need to succeed in taming cloud complexity and, ultimately, achieving AI-driven autonomous cloud operations.”