For data centre operators, their insurers, brokers and lenders
What fails in data centres, with the sources.
Every data centre fire and power failure in public reporting since 2010, written up one by one with its sources: UPS and battery rooms, backup generators, switchgear, transformers and fuel systems. Free to read. Updated as new incidents are reported.
Where the reporting names the part, UPS and battery rooms lead, then generators and switchgear. The free page counts every entry by what failed, by year and by country, and lists the ones classed as major: data centre fires and power failures.
For the underwriter and the risk engineer. A submission tells you what the broker chose to include. The register tells you what has been reported about the site, the operator and the equipment, with the source for every line, and a modelled cost and downtime range where no figure was published.
Free: what happened
Each entry: the site and operator where the source names them, what happened, the contributing factors, the lessons, every source. Never a person. A spreadsheet of the whole set arrives when you confirm an alert sign-up on the data centre page.
The record: what failed and what it cost
For subscribers, the record behind an entry adds the cause where it is established, the part that failed, the downtime and the modelled loss with a confidence band. One complete record is open to everyone.
Alerts for data centres only
One email when a newly reported data centre incident is confirmed, at most two a day, and a Monday digest of the week's data centre entries. Nothing else.
Get the next data centre fire or power failure by email
Data centre incidents only. One email when a newly reported incident is confirmed, at most two a day, plus a Monday digest. Free, no account needed, unsubscribe in one click.
By subscribing you agree to the privacy policy.
What counts as a data centre incident here. A fire, explosion, power failure, gas release or electrical injury that began in a data centre's power systems. A cloud, network, cooling or software outage with no power equipment failure is out of scope and is not recorded, so the list is not padded with outages that have nothing to do with the plant.
How the data is made. Public reports are found daily, drafted into the register's format, reviewed against fixed rules, and published as confirmed or provisional with the sources listed. Where a cause is not established in any public source, the entry says so rather than guessing. The methodology is public and the correction is free, for anyone.