Yandex Outage Disrupts Kazakhstan Taxi and Delivery Services as Recovery Begins

Services return gradually after a day of disruption

Yandex began restoring services in stages on Sunday, October 11, 2026, after an outage prevented users in Kazakhstan from booking taxis and ordering food or goods, while drivers and couriers lost access to tools needed to work. By the evening, some products were becoming available again, but the company had not announced a complete recovery or given a deadline for restoring every function.

The disruption followed reported drone attacks on three Yandex data centers in Russia over several days. Yandex confirmed that infrastructure at its Vladimir facility had been damaged in a drone attack and that the site had stopped operating. However, a direct connection between those attacks and every outage in Kazakhstan or elsewhere had not been officially established.

Kazakhstan's Ministry of Artificial Intelligence and Digital Development said servers holding personal data belonging to Kazakh users remained inside the country and were operating normally. Its explanation drew a distinction central to the incident: storing information locally does not mean every component needed to run a taxi or delivery platform is also located locally.

The effects extended beyond failed bookings. Passengers moved to competing services, inDrive reported a surge in orders and reduced its driver commission, and Dodo Pizza separately reported that its ordering systems were unavailable because of a failure at an unnamed cloud provider.

At 19:12 Kazakhstan time, monitoring systems recorded approximately 559 complaints about Yandex search and 252 about Yandex Go during the preceding hour. Those figures counted user reports, not the total number of affected people, and showed that complaints continued after recovery had begun.

Passengers, drivers and couriers lose access

Yandex Go confirmed that its services were experiencing errors or could be temporarily unavailable. Users in Kazakhstan reported being unable to book taxis, order deliveries or use some route functions. Food and goods delivery were affected alongside passenger transport.

The disruption also reached Yandex Pro, the application through which taxi drivers and couriers receive orders. Reports described the application as unavailable, leaving workers unable to continue their usual activity even where passengers still wanted to place orders.

The Yandex Go press office acknowledged the effect on both customers and workers. In a statement translated from Russian, it said:

We understand that this has affected many people's daily plans, and for drivers and couriers it means being unable to continue working as usual.

The company apologized and said technical specialists were working to restore service, while acknowledging that it could not give an exact recovery time. It promised updates for users, drivers and couriers through its official channels.

Yandex Qazaqstan separately confirmed technical difficulties affecting some services. Neither statement established that every product was unavailable throughout Kazakhstan for the same period. The reported experience ranged from errors and unstable functions to an inability to complete a booking.

Three Russian facilities affected within days

The infrastructure problems unfolded over several dates, rather than as a single reported incident. Accounts described damage in Sasovo, Kaluga and Vladimir, alongside power interruptions affecting Yandex Cloud.

• October 8: A reported drone attack was followed by a fire at the Yandex data center in Sasovo, Ryazan Oblast. The facility reportedly stopped operating.
• October 9: Drones reportedly struck the Kaluga data center, putting several modules out of operation. The equipment recovery schedule was unclear.
• October 10: Yandex reportedly added force majeure provisions to a user agreement, including military actions and certain attacks involving drones or missiles.
• Overnight into October 11: The Vladimir data center was damaged and stopped operating. Yandex confirmed the shutdown and said no one was injured.
• October 11: Service disruptions were reported in Kazakhstan and other countries. Yandex later announced a gradual restoration, with some restrictions continuing.

Yandex Cloud reported damage to the Vladimir infrastructure and warned that the remaining resource configuration was unstable. Earlier reports also described power interruptions in a cloud availability zone, a grouping of computing resources intended to operate as a distinct part of the platform.

What is confirmed about the cause?

The confirmed shutdown of the Vladimir facility provides a concrete infrastructure event behind the wider disruption. Reports also described damage at the other two sites. What has not been established publicly is the precise chain linking each damaged facility to each unavailable service in Kazakhstan.

Yandex Go Kazakhstan explained that its products depend on distributed computing infrastructure, with technical components located in different countries. A failure in one component can therefore affect functions in another country. That explanation describes how an overseas problem can spread, but it does not identify the particular module that failed for Kazakh taxi or delivery users.

Some reports presented the attacks as the cause of the Kazakhstan outage. Others explicitly said the direct connection had not been officially confirmed. The narrower conclusion is that the disruptions occurred amid confirmed and reported damage to Yandex infrastructure, while the company acknowledged dependencies across national borders.

Reports of app and payment difficulties also involved Lamoda, Dom.ru, Vkusno — i tochka and Magnit. Those complaints broadened the picture of disruption, but did not establish that all those services shared the same failure or depended on the damaged Yandex facilities.

Local data storage does not guarantee local operation

Kazakhstan's Ministry of Artificial Intelligence and Digital Development said the domestic servers used to store personal data were working normally. It also confirmed that some components required for taxi and delivery services were hosted outside Kazakhstan.

In a statement translated from Russian, the ministry explained:

Servers intended to store the personal data of users from Kazakhstan are located within the country and are operating normally. At the same time, this does not mean that the entire platform functions autonomously; certain components necessary to provide taxi and delivery services are located outside Kazakhstan.

For a passenger, the distinction is practical. A platform may retain account information on a functioning local server but still be unable to complete a ride request if another necessary computing component is unavailable. Data location and service availability address different technical tasks.

Yandex said it complied with Kazakhstan's laws and stored Kazakh users' personal data on servers within the country. Some versions of its statement also referred to jurisdictions providing appropriate protection under the country's legislation. The company did not specify which service modules depended on foreign infrastructure or which categories of information might be processed outside Kazakhstan.

The ministry's statement establishes that the local storage servers remained operational. It should not be read as a detailed audit of every data processing activity, nor does an unavailable application by itself establish that personal information has been exposed.

Recovery involves moving workloads and helping cloud customers

Yandex said it was redistributing workloads across remaining infrastructure components to reduce the likelihood of further interruptions. In simple terms, that means asking functioning parts of the system to take on work that unavailable components can no longer perform.

The company said engineers were restoring the most widely used services first. Search, Alice, Yandex Maps, taxis, mail and other products began returning, but individual functions remained restricted and technical work continued. A general announcement that products were coming back did not confirm that every function had returned for every user in Kazakhstan.

Yandex also reportedly reached agreements with VK Cloud, Selectel and K2Cloud to accelerate the transfer of data and services belonging to customers affected by the attacks. The providers assigned additional technical teams and simplified connection procedures to shorten interruptions.

That assistance concerned affected cloud customers. It did not establish that the whole Yandex Go platform had been moved to another provider or that a specific migration would restore Kazakhstan's taxi service by a stated time.

The ministry said it was coordinating with Yandex and Kazakh infrastructure organizations to support restoration. It was also working with other taxi and delivery platforms to help them remain stable under increased demand.

Demand shifts to inDrive and fares rise

The taxi disruption quickly affected a competing platform. inDrive founder Arsen Tomsky said the number of passengers and orders in Kazakhstan rose sharply that morning, upsetting the balance between demand and available drivers.

To encourage more drivers to work, inDrive reduced its commission to what Tomsky described as the minimum needed to cover mandatory taxes. No numerical commission rate was given.

inDrive uses a bargaining model: a passenger proposes a fare, and drivers may accept it or suggest another amount. When many more passengers seek rides than there are available cars, the amounts needed to secure a driver can rise.

One passenger reported that a journey usually costing about 900 tenge cost approximately 2,600 tenge during the disruption. Using those figures, the reported fare was about 2.9 times the usual amount, an increase of roughly 189%. This was an individual account, not a measure of prices across the platform, and the two fares were not independently verified.

Tomsky said inDrive would probably also limit the bargaining range during the day to restrain excessive price increases. Implementation of that proposed measure had not been confirmed by the evening of October 11. The commission reduction was announced as an action already taken; the bargaining restriction remained a possible further step.

The disruption extends beyond Kazakhstan taxis

Similar Yandex problems were reported in Russia, Belarus, Armenia, Uzbekistan, Turkey and Serbia. In Russia, complaints came from Moscow, St. Petersburg, Perm, Krasnoyarsk and Vladivostok. Both passengers and drivers reported difficulties.

The list of affected Yandex products extended beyond transport. Reports described problems with Yandex Music, Yandex Pay, Yandex Cloud, smart home services and internal tools used by employees and support staff. The breadth of those complaints pointed to difficulties across several products, without establishing that every service failed everywhere.

Dodo Pizza's website and application also became unavailable in Kazakhstan. The company said a major overnight failure at its cloud provider disrupted delivery, collection and orders placed in pizzerias. It did not name the provider, so its outage cannot be definitively attributed to Yandex.

Dodo Pizza said it was working with the provider to restore the systems but gave no recovery deadline. Customers who had paid for orders they did not receive were promised refunds after checks.

The pizza chain had previously confirmed a separate cyberattack that could have exposed names, contact details, addresses and order histories belonging to some customers. It said payment data was not affected. That earlier security incident should not be confused with the current cloud availability failure: the company's explanation did not establish a connection between them.

Deadlines and responsibilities remain unresolved

No exact date or time for complete restoration in Kazakhstan had been announced in the October 11 updates. The unresolved issues included which functions remained unavailable, how long damaged equipment would take to restore, and which overseas components were essential to the local platform.

The reported October 10 user agreement changes also require careful interpretation. Force majeure provisions concern circumstances beyond the parties' control that may affect contractual responsibility. The reported additions included military actions, terrorist acts involving drones or missiles, and the consequences of shelling and explosions.

The reports did not establish how those provisions would apply to a particular customer or worker in Kazakhstan, whether they covered all affected products, or whether any claim had been decided. The timing of a contract change is not itself a ruling on compensation.

Kazakhstan experienced a similar Yandex Go outage in November 2022, when technical problems also prevented taxi bookings. The current incident adds a documented pressure on alternatives: inDrive reported a surge in demand, while authorities coordinated with other operators to support stability under heavier loads.

The immediate test is whether service restoration reaches users consistently, rather than whether selected products have begun to return. The company had promised further updates, but no dated restoration milestone was announced.

What to Know

• Yandex taxi and delivery services in Kazakhstan suffered errors and temporary unavailability on October 11, 2026.
• Recovery began in stages, but full restoration and an exact completion time had not been announced.
• Yandex confirmed a drone attack damaged its Vladimir data center and halted operations. Damage at Sasovo and Kaluga was also reported.
• A direct link between every Kazakhstan outage and the attacks had not been officially established.
• Kazakhstan's digital ministry said local personal data servers remained operational, while some service components were hosted abroad.
• inDrive reduced driver commissions after demand surged. One reported fare rose from about 900 to 2,600 tenge, but that example was not platform pricing data.
• Dodo Pizza reported a separate failure at an unnamed cloud provider and promised checks and refunds for paid orders that were not received.