Regarding streaming replication, these days you might want to also consider a “disaggregated storage-compute” setup, like AWS Aurora, or Neon (which is open-sou

Regarding streaming replication, these days you might want to also consider a “disaggregated storage-compute” setup, like AWS Aurora, or Neon (which is open-source I believe). Re partial indexes - those are great, though you have to be careful to keep them updated as your queries evolve. E.g. if you have a partial index on a set of enum values, you should make sure to check and potentially update it whenever you add a new enum variant.