Fixing a 10-Year Performance Bottleneck

When I joined In-Play Sports, the platform had severe performance issues and many pages took 30-40 seconds to load. That was a direct usability problem for analysts who needed fast video access and real-time tagging during live matches.

Initial investigation focused on the usual suspects:

• -Database queries
• -API logic
• -Infrastructure configuration

Several smaller optimizations brought load times below 10 seconds, but the true root cause was still hiding underneath the system.

The major bottleneck turned out to be infrastructure, not application code. The production database had been running on a single CPU core for more than 10 years.

That physical limit throttled query performance under load, so no amount of application tuning could fully solve the slowdown until the infrastructure was corrected.

Once the issue was identified, the database configuration was updated so it could use multiple cores. Combined with the query optimization and caching work already in place, that unlocked a dramatic performance improvement.

• -Page load times reduced from 40 seconds to near-instant
• -Analytics responsiveness improved significantly
• -The platform could scale to thousands of teams and nearly 10,000 users

Legacy systems often hide their biggest problems in unexpected places. Sometimes the best optimization work comes from understanding the entire stack, from hardware and infrastructure all the way up to the application layer.