Equipment failure during maintenance cuts power to two Equinix SG1 data halls

What happened

On 31 May 2025 two data halls at the Equinix SG1 data centre at 20 Ayer Rajah Crescent, Singapore, lost power during scheduled maintenance. The work required a reduction in power redundancy. Equinix said that during this work it experienced an equipment failure which resulted in a power outage to the two data halls. The outage lasted more than an hour and caused widespread service issues across the country, with several internet service providers affected, including Singtel, StarHub and ViewQwest.

Equinix described the incident as minor and localised and said the issue had been resolved. The facility has 13MW of capacity. ViewQwest said the incident was triggered by an external infrastructure failure beyond its direct control and that it would work with Equinix to find the root cause.

Contributing factors

Scheduled maintenance had reduced power redundancy. The source does not identify the failed equipment and does not say how the reduced redundancy affected the outage. ViewQwest attributed its own service loss to an external infrastructure failure, a claim not confirmed in Equinix's statement.

Lessons learned

1. Before reducing power redundancy for maintenance, confirm that the remaining power paths are healthy and that a rollback plan is ready.
2. Treat any planned period of reduced redundancy as a higher-risk state, with tighter monitoring and clear limits on how long it lasts.
3. Downstream customers and network providers should be told in advance when maintenance may affect the availability of their services, and root cause findings should be shared with them.

More lessons on maintenance, inspection and ageing assets, detection and monitoring, outages, backup supply and restoration, from every incident on the register.

Source

datacenterdynamics.com