Editor’s note: This is a sober operational readout, not an outrage piece. We cover the timeline, what customers should do during incidents, and how banks can raise resilience.
The timeline
On Mar 8, 2025, DBS suffered an overnight disruption affecting mobile banking, ATMs and some NETS payments before recovery. On Jun 2, 2025, customers again reported being unable to log in for more than two hours. These episodes came after MAS in 2024 allowed earlier curbs to lapse, signaling that DBS had made progress on remediation following the 2022–2023 incidents. The 2025 outages suggest large-scale digital platforms still face complexity and dependency risks that take multiple cycles to harden.
Why it matters
Digital banking resilience has three hard problems. First, capacity and failover: can traffic be shifted without triggering hidden single points of failure? Second, change management: are configuration drift and version rollouts controlled across many teams and vendors? Third, exercise quality: do drills simulate the messy real pathways—customer authentication, inbound transfers, transaction signing and notifications—rather than just backend components? Regulators now expect banks to demonstrate not only root-cause fixes but also scenario coverage and learnings that propagate across systems and suppliers.
What customers should do in an outage
Have a practical playbook. Keep alternative payment options (a second card or another bank’s app). Use any “lock card/kill switch” function if you see unusual activity; it preserves your position while systems stabilize. After service resumes, check transaction history and bank advisories: fee waivers, reversal windows and remediation steps are typically posted alongside an incident note. If you rely on scheduled transfers or payroll, confirm execution and set alerts for missed events.
What banks should do next
Beyond adding capacity, publish clearer incident post-mortems with categories (e.g., external dependency fault vs. internal change) and specific mitigation actions. Expand observability at user-journey level (for example, login→balance→transfer) so detection is faster and business impact is quantifiable. Harmonize fallback paths for essential services like FAST/PayNow so that partial functionality can be preserved. Finally, rehearse customer communications: concise alerts, realistic restoration windows, and what users should and should not do while systems are degraded.
Bottom line
Resilience is a program, not a patch. The 2025 incidents will keep pressure on DBS—and the industry—to prove that capacity planning, failover choreography, and user-journey monitoring are routine disciplines, not emergency projects.
Sources (plain text):
CNA on Mar 8, 2025 outage; Bloomberg report on the same date; The Straits Times on Jun 2, 2025; Fintech News Singapore round-up (Jun 3, 2025).
About the Author
Marks Toms – Editor-in-Chief
Marks oversees editorial policy, compliance, and fact-checking at bankaccountsopen.
Read more articles