Comment by ropable

I have a colleague who (inexplicably) doesn't trust Postgres for "high performance" applications. He needed a database of shared state for a variable number of running containers to manage a queue, so he decided to implement his own bespoke file-based database, using shared disk. Lo and behold, during the first big (well-anticipated) high-demand event, that system absolutely crawled. It ran, but it was a total bottleneck during two days of high demand. I, who has made a New Years resolution to no longer spend political capital on things that I can't change, looked on with a keen degree of schadenfreude.

Just. Use. Postgres.