Archived Content

The following content is from an older version of this website, and may not display correctly.

Update (1 Dec): Read our follow-up report "The outage that wasn’t: Equinix Singapore" as Equinix Singapore explains what really happened.

Some customers at Equinix suffered a brief downtime over the weekend as power to a number of server racks was unexpectedly cut. This took place during a scheduled power maintenance event at Equinix’s SG1 data center in Singapore, and it is understood that power from a redundant source that was supposed to kick in failed to do so.

The planned maintenance was supposed to rectify a “thermal anomaly” with an existing transformer, and was scheduled to take two hours.

The outage appears to be limited to some customers located at level three of Equinix’s data center located at Ayer Rajah, according to a source who told DatacenterDynamics FOCUS that he had a few racks of servers go dark. When he went down after midnight to investigate, he found he wasn’t the only one who was caught by surprise, as the ground level car park was “full” of cars from other affected customers.

The source noted that this was the first time in over six years that he had encountered a power outage at Equinix Singapore. Power was restored within the two hours window, and there have been no further problems since.

We reached out to Equinix about the outage, and Clement Goh, managing director, Equinix South Asia wrote the following in an email response.

“The Equinix SG1 data center was scheduled for maintenance on Sunday 23 November between 00:01-02:00. As a result, some customers might have taken longer to recover operations from their deployment in SG1, depending on the type of service they are subscribed to.”

“In order to give our customers ample lead time before the maintenance, email notifications were sent on 22 October, one month prior and on 23 November, an hour before the maintenance was scheduled to take place.”

“In addition, we ensured that all our customers were back online and fully operational once the maintenance work was completed. We are fully aware of the importance of 24/7 access to applications and services and are committed to adhering to our SLA of >99.99% availability for all customers.”