Annual Outage Analysis 2023 - report landing page

Overview

This is Uptime Institute's annual outage analysis, published as Keynote Report 92, and the 2023 edition is presented here through its public landing page. The study merges Uptime's own survey responses with publicly reported outage records to describe how service interruptions are changing across the data centre and digital infrastructure sector. The full report sits behind a registration form or an Uptime Intelligence subscription, so this summary relies only on the key insights stated on the open page.

The first caution on the page concerns method: the ways of tracking outage frequency, severity and cost carry uncertainty, so the data should be read sceptically. Because outage reporting depends on voluntary disclosure and public visibility, published rates should not be treated as precise probabilities. For engineers who carry such numbers into design decisions, this caveat may be the most important message in the whole report.

According to the findings, outage rates are falling gradually. Most sites experienced at least one outage over a three-year window, but only a small share were serious or severe. Cost moves in the opposite direction, however: more than two-thirds of outages cost over $100,000. The number of incidents is easing while the financial price of each incident keeps rising. Another trend the report highlights is the growing role of third-party providers. Cloud, colocation, telecom and hosting companies account for a growing proportion of all outages. The page also stresses that human and management failures remain a significant factor. Reliability is therefore framed as a matter of operating processes and supply-chain dependencies as much as of equipment redundancy.

Key points

1. Methods for tracking outage frequency, severity and cost carry uncertainty, so outage data should be treated sceptically.
2. Outage rates are falling gradually.
3. Most sites had at least one outage in three years, but few were serious or severe.
4. More than two-thirds of outages cost more than $100,000.
5. Third-party cloud, colocation, telecom and hosting providers account for a growing share of outages.
6. Human and management failures remain a significant cause.
7. The report combines the Uptime survey with public outage data; the full text needs registration or a subscription.

Using it in design

• Present outage statistics in the reliability report with uncertainty caveats rather than as precise probabilities.
• In outage-cost estimates, use the finding that most outages exceed $100,000 as a reference threshold.
• Add third-party cloud, colocation and telecom dependencies to the risk assessment as separate line items.
• Treat human and management failures as seriously as equipment redundancy and reflect them in operating-procedure checklists.

Who should read it

Engineers and managers responsible for data centre reliability, risk and operations decisions.

Publication details

Authors

Uptime Institute; A. Lawrence; L. Simon

Publisher

Uptime Institute

Year

2023

Document number

UII Keynote Report 92

Document language

English

Source type

Industry body

Access

Citation only

Licence

Copyright retained; citation only — the publisher’s terms apply to the original document

Citation