Designed for the waves – how we absorb offline spikes in retail payments
A recent network outage in the Nordics pushed many payment terminals into offline mode – and when the connection came back, thousands of queued transactions hit all at once. For us, it became a live test of how our platform handles sudden, uneven traffic.
The result? Peaks of nearly 300 transactions per second over several hours, with no downtime or impact on our services.
"Resilience is both engineering and readiness. You need strong architecture, but also people watching the right things."
Payments don’t happen in isolation. Terminals, network operators and central services all depend on each other, and when one of them has a bad day, everyone feels it.
“Networks rarely recover in a straight line – they come back in waves. That’s why we design for the peaks, not the averages,” says Kristian Lingsom, Head of IT and Security at Swedbank Pay.
Resilience in practice
What made the difference was not just capacity, but preparation. Our systems automatically scaled to absorb the surge, while our teams monitored the flow live, tracking patterns and thresholds so we could intervene if needed. That combination – automated elasticity with human oversight – is central to how we build resilience into the core of our platform.
“Resilience is both engineering and readiness. You need strong architecture, but also people watching the right things,” Lingsom adds. “That's how we stay calm when the system is anything but – because we’ve already planned for the moments that don’t follow the script.”
Build for the peaks – not the everyday
The incident reinforced something we plan for every day: traffic isn’t linear. It doesn’t follow nice averages. It comes in bursts, especially after disruptions in the surrounding ecosystem. Building for resilience means optimizing for the extreme moments, not the comfortable ones.
For our customers, those moments should feel like nothing at all. Payments should simply work.
And in this case, they did.
