Interesting timing given that their CI/CD pipelines are currently down and have been for hours.
Damn, it seems a few of their (outages|degradation) centers around poor index management. I would think PG has better tooling to support surfacing that silliness, but I freely admit I'm not a DBA