Is Tornado Datacenter down?

We're currently monitoring a outage of the public power grid connection for TornadoDC FRA1. The datacenter has switched over to generator power supply and is supplied from there, UPSes took over the load until generator load transfer. We're investigating the issue.

We had a defective fuse in the building NSHV. The same has been exchanged, the facility has switched back to regular grid supply.

The incident was caused by the failure of an NH fuse within the building’s main low-voltage switchboard \(NSHV\), specifically within the section supplying Tornado Datacenter. To mitigate the impact of such upstream power failures, Tornado Datacenter operates an automatic standby generator system that continuously monitors the quality and availability of the public grid supply. If grid parameters fall outside the defined operating limits, the generator starts automatically. Once generator voltage and frequency are stable and within specification, the datacenter’s electrical infrastructure automatically transfers the load to generator supply. The complete process takes approximately 20 seconds. During this transition period, the datacenter load is supplied without interruption by multiple UPS systems operating from battery power. The entire emergency power chain is tested every three months as part of scheduled black-building simulations. During these tests, the public grid supply to the datacenter is intentionally disconnected for approximately one hour to verify the operation of the UPS systems, automatic generator start, load transfer, sustained generator operation, and subsequent return to utility power. During the incident in question, the public grid supply to the datacenter became unavailable following the failure of the upstream NH fuse. As designed, the UPS systems immediately maintained power to the datacenter while the standby generator started and assumed the load. Qualified in-house electrical engineers subsequently identified and replaced the failed fuse. After the building supply had been restored and verified, the datacenter infrastructure automatically transferred back from generator operation to the public grid. The underlying cause was an increased electrical load associated primarily with higher cooling demand and the rapid growth of Tornado Datacenter. Several heatwaves during summer 2026, combined with the operation of the newly implemented chilled-water cooling system, resulted in significantly higher peak loads. Over a period of several months, repeated operation close to the fuse’s thermal limits led to progressive thermal ageing and ultimately caused the fuse to operate during a peak-load condition. As a long-term capacity expansion, we are currently planning the implementation of a new transformer substation, which is scheduled to become operational in the first quarter of 2027. This expansion will substantially increase the available electrical capacity and provide additional headroom for the continued growth of the datacenter, especially capacity in future Room B \(~1MW\).

TornadoDC is currently monitoring increased cooling demand caused by elevated solar heat gain. Due to the ongoing roof modernization above the datacenter, the roof insulation has been temporarily removed, resulting in a significantly higher thermal load on the facility and increased power consumption by the cooling infrastructure. As a precautionary measure, we have activated additional cooling capacity to maintain stable operating conditions. We have also ordered additional equipment to further strengthen the site's cooling capabilities during the remaining construction period. The measured temperature within the cold aisle remains well within our SLA-guaranteed limits, and there is no impact on customer services. The building owner and the construction contractor have been informed of the situation. We have received confirmation that the affected roof section above the datacenter has been prioritized and that work on this area will be completed as quickly as possible. We will continue to closely monitor the situation and provide further updates if necessary.

We got informed by the building owner, the contractor responsible for the roof modernization will start the works as of today in order to restore outer insulation above the datacenter.

In coordination with the manufacturer of the chilled-water system, we have implemented several optimizations to maintain more stable temperatures while reducing capacity fluctuations. Temporary roof insulation has also been installed to minimize additional heat gain. Additional cooling equipment is scheduled for delivery tomorrow, providing increased thermal capacity and additional headroom for the remainder of the 2026 summer season. This will allow the existing DX air-conditioning units to remain reserved for emergency and exceptional operating conditions, such as the current situation, rather than being required during normal operation. In summer 2027, the existing DX air-conditioning units will be replaced with in-row CRAH units, further increasing cooling capacity and improving overall efficiency and temperature stability.

We're closing the same, cold aisle temperatures were stable over the past days and behaved as planned.

Due to a electrical problem, cooling has been impacted in Rackroom A of TornadoDC FRA1. We're currently working on restoring the same, mid the incident we have deployed backup cooling units to tackle the temperature. On-Duty personnel is working to reduce temperatures on a normal level.

Temperatures are normalizing into SLA range levels. We're monitoring the infrastructure and investigating it's state.

Temperatures in the cold aisle have returned to earlier levels, currently at around 22-24°C. On-Duty personnel is on-site to keep monitoring operations and make adjustments if necessary.

This incident has been resolved.

At 18:31 CEST, the on-call engineer was paged due to rapidly increasing temperatures across Rack Room A. Upon receiving the alerts, the incident was escalated internally at 18:33 CEST. Simultaneously, the on-call engineer departed for the datacenter to respond on-site. Following the escalation, additional team members joined the incident response. As an immediate mitigation measure, backup air conditioning units were activated at 18:34 CEST. Due to the elevated ambient temperatures, the firewall providing remote access to the datacenter infrastructure shut down automatically. At 18:39 CEST, the on-duty engineer arrived on-site and verified the availability of electrical power and the chilled water generation system. Both systems were operating normally and showed no indications of failure. At 18:43 CEST, the on-duty engineer identified an electrical fault affecting the cold-air circulation within the cold aisle of Rack Room A. Power to the free-cooling infrastructure was restored immediately. In addition, the chilled water supply temperature was reduced to increase cooling capacity and accelerate the recovery of room temperatures. We will use the findings from this incident to further strengthen the resilience of our infrastructure. The identified lessons learned will be incorporated into our engineering, monitoring, and operational procedures to reduce the likelihood and impact of similar incidents in the future.

We're currently closely monitoring the ongoing heatwave across Europe, with daytime temperatures exceeding 40°C. To ensure business continuity, we implemented several measures in advance: - The datacenter is cooled using a chilled water system, which has been verified to operate within the manufacturer's specified ambient temperature range - Adiabatic pre-cooling has been added to the chilled water system to reduce the temperature of incoming air before it reaches the mechanical cooling equipment - The datacenter is currently operating in recirculation mode. Outside-air free-cooling dampers are closed to maximize cooling efficiency during the heatwave. - Grid power quality is continuously monitored, and standby emergency generators are available to seamlessly take over the electrical load if required - Our datacenter on-call engineers are continuously monitoring operational metrics, with particular attention during peak temperature periods We do not expect operational instability following implementations made since 2025.

We're closing the same, since the infrastructure dealt with the heatwave flawlessly. Outside temperatures were hitting 41°C past days.