An Alarming Increase in Cloud Outages
In 2024, Parametrix, a leader in digital business interruption risk solutions, observed an 18% rise in critical cloud service disruptions among the top three public cloud providers. These disruptions not only increased in frequency but also in duration, with outage times expanding by 18.7% compared to the previous year.
Escalating Risks and Duration
The Cloud Outage Risk Report 2024 sheds light on the escalating risks tied to cloud infrastructure, which businesses increasingly depend upon. A concerning trend emerged: six major outages in 2024 each lasted over 10 hours, culminating in nearly 100 hours of total downtime. While North America bore the brunt of these interruptions, Europe and Asia were also affected.
Human Error: A Leading Cause
Human error continued to dominate as the primary cause of these disruptions, responsible for 68% of the incidents, a notable rise from 53% the previous year. This growing dependency on cloud computing highlights the systemic risks businesses face, especially as the demand for cloud services surges, powered by the advancement of generative AI and other cutting-edge technologies.
CEO Insights on Cloud Reliance
“The critical role of cloud infrastructure in global business operations was emphasized in 2024,” remarked Jonathan Hatzor, Co-Founder and CEO of Parametrix. He cited major incidents like the global CrowdStrike outage in July, AWS’s service disruption in the US-EAST-1 region, and Google Cloud’s power failure in Frankfurt as examples of the growing systemic risks faced by enterprises reliant on digital supply chains.
Investment and Risks
Hatzor further noted, “The swift adoption of generative AI has intensified demand for cloud services. In the third quarter of 2024 alone, the top three cloud providers invested $82 billion in infrastructure. This growth, while driving innovation, also amplifies the risks of cloud outages, which can disrupt operations, erode customer trust, and incur significant financial losses.”
Regional Impact and Provider Performance
Despite the heavy investment in infrastructure, the rapid growth of these services introduces new risks. Google Cloud faced a striking 57% increase in critical downtime, whereas Microsoft Azure experienced a reduction of over 20%. Yet, AWS stood out as the most reliable, with only two critical outages, each lasting under 30 minutes.
Data-Driven Risk Modelling
Parametrix’s data-driven approach to cloud risk modelling offers invaluable insights for insurers and reinsurers. The report stresses the importance of understanding regional and service-specific outages, alongside the patterns of attritional risks that could escalate into catastrophic events.
The Necessity for Nuanced Models
Sharon Haran, Chief Commercial Officer of Parametrix, emphasized the need for more nuanced models to appraise cloud outage risks. “Our third annual report underscores a crucial lesson: cloud risk cannot be viewed as a single systemic event. It is vital to comprehend regional and service-specific outages, attritional risk patterns, and escalation pathways to define realistic disaster scenarios. Models lacking these elements may misrepresent exposure and lead to inaccurate risk assessments.”
Adapting to the Evolving Cloud Landscape
Haran added, “The rapid expansion of cloud services, driven by AI adoption and digital transformation, has introduced new layers of risk requiring a more nuanced, data-driven approach to modelling. Cloud failures can no longer be perceived as rare, systemic events. They must be assessed with precision, acknowledging both attritional and catastrophic scenarios. Our models continue to aid insurers in defining realistic disaster scenarios and ensuring that coverage structures accurately reflect the true nature of cloud outage risks.”
Urgent Need for Accurate Risk Management
The report underlines that as the frequency and duration of critical cloud outages rise, there is an urgent need for more precise risk management strategies. Businesses, insurers, and service providers must adapt to these evolving challenges by employing data-driven models that consider the complexity and scale of cloud service interruptions.
