Observability without AIOps is just noise.
Moogsoft CEO Phil Tee delivered the proverbial good news / bad news combo to those who attended his presentation at AWS re:Invent 2019 in Las Vegas.
First the bad news: statistical analysis of IT metrics data is “stuck in the stone age,” hampering IT Ops and DevOps teams’ ability to detect and fix problems that impact digital services.
And the good news? Applying AI and ML algorithms is equally as relevant to time series metrics as it is to other areas of AIOps.
“Our belief is that the combination of AI, metrics and logs is where we need to be thinking,” Phil told the audience.
The complexity problem
As IT environments get increasingly hybrid, ephemeral and distributed to support an organization’s digital transformation efforts, they become correspondingly more complex and harder to monitor.
“The ‘software-defined everything’ world is incredibly changeable and almost impossible to pin down in terms of its configuration,” Phil said. “There is practically no ‘steady state.’”
While IT operators struggle to maintain the quality, stability and performance of services, customers get more demanding every day, expecting zero downtime and latency.
“There’s an avalanche of complexity out there which we’ve got to collectively tame,” he said.
He cited results from a recent survey in which a majority of CIOs reported troubling trends that point to “systematic failures” in service assurance, including that:
- Customers notice service incidents before their IT support team does
- Existing monitoring solutions detect less than half of performance issues and outages
- Growing IT complexity leads to more outages
What to do?
Phil, a serial innovator who has devoted 25-plus years to improving IT operational management, co-founded Moogsoft in 2011 to apply AI and ML to this problem, after noticing legacy rules-based products couldn’t handle the scale, speed and dynamism of modern environments.
At first, most questioned this approach, but Phil has been proven right. Moogsoft is a leader in the hot, new market for AIOps (Artificial Intelligence for IT Operations), which is now acknowledged as truly transformative for IT Ops and DevOps teams.
More than 130 companies, including 20 of the Fortune 1000, use the Moogsoft AIOps platform to streamline IT operations, quickly detect and fix incidents, prevent outages, meet SLAs, and boost their digital transformation.
Most recently, Moogsoft has tackled the monitoring of metrics data, a key feature in their AIOps platform and available in the Express entry level package, now in beta. Compared to logs, a key difference when monitoring metrics is the volume of data. An at-scale environment may generate thousands of log messages per second, while producing hundreds of thousands of metrics per second, Phil explained.
Observability without AIOps is just noise
But all that raw metrics data by itself contains little information because it lacks context. Knowing that the CPU usage on a server is at 94% is of little value if you don’t know if it indicates normal functioning or a potential problem.
“You need to know a lot more, like: What was it like yesterday? What was it when the system was doing something different? Do I have only one of those servers or is it part of a 1000-server farm?,” Phil said.
This is where AIOps comes in. Moogsoft ingests monitoring data, and applies AI and ML algorithms to it, eliminating noise, detecting anomalies, correlating relevant metrics, alerts and log events, surfacing them as contextualized incidents and identifying problem root causes.
The bottom line is giving IT operators and DevOps teams more visibility and a better handle on the true state of their digital infrastructure — true observability — so that they can maintain service quality by fixing the right incidents all the time on a timely fashion, Phil said.
“AIOps is the path to continuous assurance, the path to making sure there’s 100% service availability,” Phil said.
Watch the full video of Phil’s AWS re:Invent 2019 presentation for a lot more details about these and other topics related to AIOps trends and to Moogsoft’s technology and products.