It is very rare to see a Jepsen report that concludes with a note that a project is being too humble about their consistency promises.
Finding effectively only a single obscure and now fixed issue where real-world consistency did not match the promised consistency is pretty impressive.
> Finding effectively only a single obscure and now fixed issue where real-world consistency did not match the promised consistency is pretty impressive.
They also admitted, that testing framework cannot evaluate more complex scenarios with subqueries, aggregates and predicates. So it is possible, that PG consistency promises are spot on or maybe even overpromising.
Strong consistency is easier when throughput and concurrency are limited.
Zookeeper is impressive in many ways. However, unless something changed drastically in the last two years, Zookeper throughput will always limit it to configuration/metadata/control-plane rather than a primary/data-plane use cases.
Finding effectively only a single obscure and now fixed issue where real-world consistency did not match the promised consistency is pretty impressive.