The average cost of IT outages has increased to $1.9 million per hour, with the median downtime for a high-impact outage being 77 hours.
This is according to observability company New Relic’s newly released 2024 Observability Forecast, which surveyed 1,700 IT professionals from 16 countries.
Just a decade ago, Gartner reported that the average cost was $300,000 per hour, so these new numbers from New Relic mark more than a 500% increase in the last 10 years.
Further, IT teams are spending about 30% of their time — or 12 hours in their 40 hour work week — dealing with service interruptions. Within the past two years, 35% of outages were the result of network failures, 29% were because of a third-party or cloud provider outage, and 28% were due to human error.
The practices that are most likely to reduce downtime include root cause analysis and post-incident reviews (37%); monitoring DORA metrics (34%); tracking latency, traffic, errors, and saturation; and managing Mean Time to Detect and Resolve outages (33%).
AI adoption has also had an impact on the need for observability, with 41% of respondents saying AI and an increased focus on security, governance, and risk are driving them to need observability more.
Another finding of the report is that full-stack observability can reduce downtime by 79%, or 70 hours compared to 338 hours per year. This can save companies up to $42 million and reduce hourly outage costs by up to 48%.
Other findings were that observability has a 4x ROI, companies prefer to consolidate their observability to a single platform, and over half of respondents are using at least one open source observability solution.
“This year’s Observability Forecast shows the real world benefits of implementing full-stack and business observability,” said Nic Benders, chief technology strategist at New Relic. “The companies that embrace those practices have less downtime, fewer critical outages, and a higher ROI. This will be even more true as more organizations increasingly adopt AI in their business, which can lead to huge costs and reliability issues.”
