Re:Index bloat reduced in PostgreSQL v14
by marcosdumay on Saturday September 11, 2021 @09:01AM
(#28491643)
Well, 32MB * 400 clients ~= 12GB. It's not an irrelevant amount, but if you have that much load on your DB, you should have a beefy machine anyway, and that for one of the 2 main memorysets isn't any extravagant (and the shared memory is much smaller). In fact, I wouldn't even try to run SQL Server on that amount (not to talk about Oracle).
But yeah, if you go full microservices automatically deployed over kubermenets all connecting to the same place, you will need something in between so your DB server doesn't get crazy. This setup would break any central DB anyway, it's just that different DBMS would break for different reasons.
But yeah, if you go full microservices automatically deployed over kubermenets all connecting to the same place, you will need something in between so your DB server doesn't get crazy. This setup would break any central DB anyway, it's just that different DBMS would break for different reasons.