Yandex Partially Restores Smart Home After Boiler and Smart Plug Automations Fail

On October 13, 2026, Yandex partially restored its smart home platform. Before that, users had reported automation failures across four classes of devices at once: water heaters, baby monitors, gas and electric boilers, and smart plugs. The most alarming case came from one home, where the automatic boiler shutoff triggered by a temperature sensor did not fire overnight. No one was hurt.

What is known so far

Public details are scarce. In the available materials, the company does not name the cause of the outage or the number of affected homes. The picture comes from user complaints and one specific incident.

• Date: October 13, 2026.
• Affected categories: water heaters, baby monitors, boilers, smart plugs.
• Type of failure: automations did not fire, meaning "if condition, then action" scenarios.
• Status: service partially restored; full recovery is not confirmed.
• Most serious incident: overnight, a boiler did not shut off on its temperature sensor's signal; no one was hurt.
• Not disclosed: the root cause, the scale, the timeline for full recovery.

Why a scenario failure is more dangerous than an assistant failure

When a voice command fails, the person hears an error and repeats the request. The failure is visible immediately. Automation works differently: it has to fire when nobody is around, most often at night or while the owners are away. If a scenario stays silent, the user finds out from the consequences. That is why shutting off a boiler on temperature, switching off a plug with a heater attached, and monitoring a baby monitor are functions where a silent failure costs more than it does for any "smart" feature like turning on a light.

An important technical question is where the scenario logic ran. If it ran in the cloud, an outage on the provider's side becomes a failure for all users at once, even when the sensor, the boiler and the home network are working fine. If it ran on a local hub, the home loop should have kept working. The material does not specify this, so no firm conclusions can be drawn. But the set of affected devices itself shows that some automations depended on the central platform.

There is also an industry-level angle. When a single vendor controls accounts, the device cloud, scenarios and the voice assistant, a failure in any link hits the whole chain. For mass-market consumer systems, this is the usual architectural price of convenience. The news shows that this price is measured in more than a lamp that will not turn on.

What smart home owners should do

The main rule: protection against emergencies should not be left to a cloud scenario alone. The practical steps are these.

• For boilers and water heaters, rely on the standard hardware safeguards: built-in thermostats, safety valves, thermal relays. A smart scenario should be an add-on.
• Do not connect heaters or other high-power loads through smart plugs without independent overheat protection. Mechanical timers and thermal cutoffs suit such tasks.
• After the partial recovery, open the app and check that your scenarios are enabled and actually fire. The outage may have left some automations inactive.
• Turn on notifications for device and hub unavailability, so you learn about a failure before nightfall rather than in the morning.
• Where possible, move critical rules to a local hub that runs scenarios without contacting the cloud. Test this by disconnecting the internet at the router for a few minutes.
• For a baby monitor, keep a backup monitoring channel, for example an ordinary audio baby monitor with no cloud connection.

Users have a right to expect an incident analysis from Yandex: what broke, how long the outage lasted and which scenarios run locally. Until such data is published, it is wiser to treat any automation responsible for safety as dependent on an external service and to back it up with hardware.