An Indian UPI payments app is handling about 12,000 payment requests per second at 20:05 IST on a festival evening. Success rate falls from 98.6 to 71 percent within three minutes. Twenty engineers have already joined a chat channel, two VPs are asking for updates, and the customer support queue is filling up.
The interviewer asks“Payment success has just dropped to 71 percent on the biggest festival evening of the year and you are paged as incident commander. Run the room for me: what do you set up, who does what, and walk me through the first thirty minutes minute by minute.”
I declare a SEV1 immediately, take the IC role and stay off the keyboard. In the first five minutes I name an ops lead to drive the technical work, a comms lead for executives, support and status updates, a scribe for the timeline and decision log, and a bank liaison for NPCI and partner banks, and I move everyone onto one channel and one bridge. I separate mitigation from diagnosis: pre-approved levers like rollback, feature shedding or steering away from a failing bank get pulled in parallel with timeboxed hypotheses. Updates go out every 15 minutes whether or not anything changed, executives get a dedicated channel and a liaison, and we stand down only against explicit criteria: success rate back at baseline for 30 minutes at peak volume and in-doubt transactions handed to reconciliation.
A free account unlocks the rest of it, and saves your progress across every track.