Postgres 18 just dropped with a bunch of nifty additions—think UUIDv7 for time-ordered unique IDs, an “over explain” module to dig deeper into query plans, and composite index skip scans for faster lookups. But the real headliner is asynchronous IO: Hussein breaks down why it matters, how Postgres went from purely blocking disk calls to spinning up worker threads or even leveraging io_uring, all while keeping a synchronous fallback around.
He also walks through the real-world challenges of shipping async IO, shows off the new io_method options (worker, io_uring, sync), and teases what’s still on the roadmap—better backend writer support, direct IO, and tweaking the worker mode for peak performance. If you care about squeezing every bit of speed out of your database, this episode is a must-watch.
Watch on YouTube
Top comments (0)