Storing data locally doesn't keep your service local

We already wrote about the Yandex outage that stopped taxis and cloud services in Kazakhstan; this note covers what came after. On the evening of 9 October, ordering a ride through Yandex Go failed in Astana, Almaty and Aktobe, and DownDetector logged more than 2,900 complaints within an hour. The cause was trouble at the company's data centers outside Kazakhstan, while the press service called the failures isolated and said they were fixed quickly.

Almost at the same time, Yandex reminded users that Kazakh citizens' personal data is stored in Kazakhstan. The bluescreen piece examines what the localization law actually requires, and the key point is the gap between two things: where data is kept and where the service itself runs. Data can sit at home while processing and service logic depend on sites abroad, so a failure there still stops work here.

Ask each critical vendor — CRM, accounting, AI tools — two questions: where is our data stored, and where does the system physically run? Get the answer in writing and find out whether you have a manual fallback if a foreign site goes down. A protected environment and role-based access are where any AI project should start; in 1Cifer each employee's agent sees only what their position allows.