The Challenge
An automated point-expiry system had run flawlessly in production for years. Then a new customer arrived with extremely high transaction volumes — far beyond anything the system had encountered — and that scale exposed critical bottlenecks in the underlying queries and data-access patterns. What had been invisible at normal volumes now threatened to bring production down.
The pressure was real: this was a live system with real customers, and the automation could not simply be switched off while it was fixed.
The Approach
The work had to happen carefully, on a live system, without disrupting the automation that was still running.
- Analysed query and data-access patterns under load to find exactly where the volume was causing the system to choke, rather than guessing.
- Applied targeted optimizations to the hot paths — improving how data was queried and accessed — without touching or destabilising the live automation.
- Ran a custom off-hours program to clear the accumulated backlog of data that had built up, doing the heavy cleanup outside business hours so daytime operations were never affected.
Because the fixes were surgical and scoped to the real bottlenecks, there was no need for a risky rewrite or extended downtime.
The Result
The crisis was resolved within half a day, with zero business disruption. The point-expiry automation kept running throughout, the new high-volume customer was supported, and the system was left healthier than before — proof that a production emergency doesn't always mean a rebuild; often it means finding the precise bottleneck and fixing exactly that.