Incident summary — Degraded performance and delayed order visibility
Period: 22 August 2026 from approx. 23:00 CEST until 23 August 2026 at 13:12 CEST
Affected services: Viskan Backoffice, Viskan API, Viskan Automations
Impact
From late Saturday evening, response times were degraded across Backoffice, the API and automations. During Sunday morning the situation worsened, and web orders began queuing before becoming visible in Backoffice and via the API. Orders were received and processed correctly throughout; none were lost or dropped. The backlog cleared during the morning and normal performance was restored at 13:12 CEST.
Cause
The incident was caused by insufficient capacity headroom in one part of our platform. As we have gradually moved more customers onto this environment, its capacity has not been scaled at the same pace, leaving a smaller buffer than we consider acceptable. From Saturday evening, an unusually high volume of automated requests from a single integration pushed load beyond what the available capacity could absorb.
What we have done
We added capacity to the affected environment, which restored performance and allowed the order queue to drain. All queued orders were processed and made available without data loss.
What we will do
Review capacity and headroom across our environments, so that growth in usage is matched by scaling in advance.
Adjust our monitoring and alert thresholds so that this type of load pattern is detected before it affects customers.
Introduce guardrails on request volumes, so that traffic from a single integration cannot affect service for others.
