Яндекс.Метрика
Postmortem

Incident postmortem

Root cause analysis and timeline of the technical outage on March 1, 2026.

Incident summary

Date and timeMar 1, 2026, 16:08
CauseOur cloud provider shut down its infrastructure because of billing issues on its side
SeverityHigh
Duration11 min of full downtime + 58 min of partial downtime
Affected systemsAPI server, mobile and desktop apps
Affected usersAll users

Incident details

Resolution timeline:

  • 16:04 – Monitoring systems detected that the service was down.
  • 16:06 – Cause identified: the provider's cloud resources were unavailable.
  • 16:08 – The provider's on-call engineers confirmed that the resources had been shut down because of billing issues. Incident logged.
  • 16:10 – First reports from users came in.
  • 16:14 – The provider restored access to the servers, and the service started to recover.
  • 16:19 – Backend systems and mobile apps were back up. We found an error that kept the desktop app client from recovering and started looking into it.
  • 17:07 – Problem identified, and we started fixing it.
  • 17:15 – The desktop client error was fixed, and the desktop apps gradually came back online.
  • 17:17 – Performance was fully restored.

Root causes:

  • Our cloud provider ran into a payment billing error, which shut down the infrastructure.
  • After access was restored, we found an error in the client side of the service that kept the desktop apps from reconnecting automatically.

Actions taken:

  • We agreed on additional safeguards with the provider in case of billing issues.
  • We fixed the error in the client side of the desktop apps.

Lessons learned and recommendations:

  1. Add a backup mechanism so the desktop client recovers automatically when the infrastructure restarts.
  2. Monitor billing status with our cloud provider to catch potential problems early.