U k r V i s t i

l o a d i n g

Power Outage Incident at De Novo Data Center Explained

Learn about the causes and consequences of the power outage incident at De Novo data center that occurred on April 26.

image

De Novo has provided insights into the first power outage in nearly 15 years, which left the Kyiv data center without electricity for 14 minutes. This disruption caused significant issues, affecting services like the "DiiA" app, "Nova Poshta", and banking terminals.

Journalists from EП visited the Kyiv data center and spoke with company management to gather details about the incident.

De Novo serves over 150 clients, including "Oschadbank", "Raiffeisen Bank", "Nova Poshta", Kernel, "Kyiv Digital", and the "DiiA" app. About 70% of clients utilize the company's cloud services, while the remaining 30% host their equipment in the data center.

On April 26, EП reported that the cause of the outage was not a cyberattack. The company identified an issue that arose during the final phase of replacing uninterruptible power supplies as the primary reason.

To ensure continuous power supply, engineers used a temporary power bypass. At 8:03 AM, as per plan, the switch was turned to the "off" position in preparation for connecting a new line. However, due to a mechanical defect, the switch did not operate fully, even though the indicator showed everything was off.

When power was supplied through both the new line and the bypass simultaneously, the automatic system interpreted this as a serious error and shut down to prevent an accident.

Power was restored within 14 minutes. However, due to the complexity of the systems, full recovery of De Novo's cloud infrastructure took until 9:45 AM, and the restoration of individual client IT systems took several more hours. Overall, the effects of the incident were resolved by 1 PM.

De Novo's CEO, Maksym Ahiev, explained that predicting such a situation was nearly impossible—the switch did not move to the required position, even though the indicator showed everything was normal.

"In over 5000 days of the data center's operation since November 2010, this is the first downtime lasting 15 minutes. Our contracts allow for up to 96 minutes of downtime per year, and continuous service provisioning, so the incident did not warrant a review of agreements," noted Ahiev.

According to him, engineers from Schneider Electric, the energy equipment manufacturer installed in the data center, promptly arrived at the site. The chief engineer of the data center, Oleksandr Yurchenko, added that this switch had been in operation since 2018.