An Indian UPI payments app routes traffic through two PSP banks to the NPCI switch. On a sale evening, transactions routed through one PSP bank start timing out at 22 percent while the other is healthy. The bank's first-line NOC replies to an email that their systems show no issue. Meanwhile in-doubt transactions are piling up and customers are complaining on social media.
The interviewer asks“Your telemetry says a partner bank is the problem, but their NOC says everything is green. How do you communicate with banks and partners during an incident, what do you send them, and what happens after it's over?”
Partner communication is built before the incident: a tested escalation matrix per bank with names, phone numbers and time-based triggers, and an agreed format for evidence. During the incident I lead with evidence, not blame: a sample of UPI transaction IDs, RRNs, millisecond timestamps, response codes and the API involved, plus aggregate rates by time, phrased as what we observe. I agree an update cadence on first contact, escalate to L2 and a joint bridge if two emails don't converge, and mitigate on our side in parallel by shifting traffic within the other bank's limits. Compliance is looped in early because RBI, CERT-In and NPCI reporting windows can be hours. Afterwards we run a joint review with aligned timelines and action items owned on both sides.
A free account unlocks the rest of it, and saves your progress across every track.