This already was de-facto the case, which made much of the logic around handling "but what if .max is nil?" somewhat tricky to reason about.
Extracted from #970 and fleshed out here.
All our staging & production clusters have VMs meeting this criteria, with the notable exception of a single, already-broken VM in dev-us-east-2-beta (it's a long story).
This already was de-facto the case, which made much of the logic around handling "but what if .max is nil?" somewhat tricky to reason about.
Extracted from #970 and fleshed out here.
All our staging & production clusters have VMs meeting this criteria, with the notable exception of a single, already-broken VM in dev-us-east-2-beta (it's a long story).
More here: https://neondb.slack.com/archives/C03TN5G758R/p1718742308251569
Checked each cluster with: