Relic research shows high impact IT outages carry $76M median year

New Relic released its 2025 Observability Forecast, the industry’s most comprehensive report on the state of observability. Surveying over 1,700 IT and engineering leaders and team members across 23 countries and 11 industries, the report highlights key focus areas, challenges, and trends influencing observability investments like the growing adoption of enterprise AI. The data showed that the cost of any digital business downtime is profound—with high-impact outages carrying a median cost of $2 million USD per hour, or approximately $33,333 for every minute systems remain down. The annual median cost of high-impact IT outages for businesses surveyed is $76 million USD per year.

Full-stack observability slashes cost of outages in half

Full-stack observability is defined as having visibility across the technology stack, organized into five key categories: infrastructure, applications and services, security monitoring, digital experience monitoring (DEM), and log management. For respondents with full-stack observability, the research shows the cost of significant IT outages is cut in half, with high-impact outages averaging $1 million per hour versus $2 million per hour for those without full-stack observability. Teams with full-stack observability also report fewer outages (just 23% experience high-impact outages at least weekly versus 40% without) and detect them 7 minutes faster with an average mean time to detection (MTTD) of 28 minutes. 

Outages distract engineers from innovating

The research also reveals the top three primary causes of these outages—network failure, third-party or cloud provider services failure, and deploying software changes. Notably, the data shows that engineers are spending 33% of their time fighting fires or addressing disruptions, taking focus away from developing new features or coding innovations. Forty-one percent of leaders reported they still learn about service interruptions through inefficient means, such as customer complaints, incident tickets, or manual checks. 

Increasing AI adoption heightens awareness of visibility challenges  

As LLM-powered applications and agentic AI become more widely adopted, they introduce new visibility challenges that traditional monitoring methods cannot solve. Without modern observability built for the AI era, silent issues can ripple through systems unnoticed. The research shows the majority of organizations have recently become aware of these challenges, as the use of observability’s AI monitoring capabilities went from 42% in 2024 to 54% in 2025—a double-digit growth rate year-over-year. Further, only 4% of businesses are not deploying or planning to deploy AI monitoring.  

AI is being used to monitor AI

Organizations adopting AI at scale require a deeper level of system insight like using AI-strengthened observability platforms, in real time, to reveal how AI models interact with pipelines, APIs, and downstream applications. When asked which AI features respondents think would improve their organization’s incident response or observability practice the most, the number one cited feature was AI-assisted troubleshooting. In order, the next most popular cited features were automatic root cause analysis, AI-assisted remediation actions, forecasting and predictive analytics, and AI-assisted generation of post-incident reviews.

"This year's Observability Forecast illustrates that amidst the growing reliance on digital services and AI coding tools, outages are costing businesses more than ever before," said New Relic CEO Ashan Willy. “Despite the promise of AI to speed application production, the data reveals engineering teams are still losing a third of their time battling issues that are difficult to pinpoint. Full-stack observability can halve the cost of a major outage while speeding up detection and resolution, freeing up teams to focus on innovation that meets business objectives."

Other key findings from Relic research shows: 

Observability reduces MTTD and delivers positive business value.
Nearly seven in ten (68%) organizations report measured improvements in MTTD since adopting observability. Additionally, the vast majority of businesses (75%) report a positive return from their observability investments, and nearly one in five (18%)
 say they are realizing 3-10X ROI.

Executives and practitioners recognize impactful benefits stemming from better and more automated observability. For leadership, reduced unplanned downtime is the top cited benefit of observability, followed by overall operational efficiency improvements, and reduced security risk. For practitioners, the top benefit is reduced alert fatigue, followed by faster troubleshooting and root cause analysis, and improved collaboration across teams.

Businesses are taking steps to consolidate observability tools. The data shows a 27%
decrease in the average number of observability tools per organization from 2023 to 2025. The median number of tools used today is four tools per organization. 52% percent of organizations plan to consolidate observability tools onto a unified platform in the next 12-24 months, with 48% also increasing investment in AIOps/ML capabilities.