We realised our legacy code didn’t need a rewrite — it needed a performance plan.
This talk shares how we tackled real production issues like memory bloat, slow queries, and Sidekiq queues piling up under load — where simple fixes like batching jobs or tweaking retries didn’t help much.

We thought ActiveRecord was the bottleneck, but fixing it only scratched the surface.
We tuned the garbage collector, optimised query access patterns, chose more efficient data structures, restructured background jobs, and introduced smarter database caching and indexing strategies.

I’ll share how we analysed query plans, fine-tuned database performance, and replaced slow, fragile logic with leaner, more explicit code paths that scaled better and failed less and Rails stayed right at the centre of it all.