The headless service pods should always be discoverable even if they're not ready, as it's used for cluster joining. However, vault, vault-active, and vault-standby services should be able to set publishNotReadyAddresses to false to ensure client requests are always routed to a ready pod. Before this PR, all services used the same values.yaml setting, so it wasn't possible to have different values for the different services. Ideally, it would be nice to set the default publishNotReadyAddresses value to be false at some point, but I just wanted to keep this PR to the minimum possible (minor) breaking change.
The headless service pods should always be discoverable even if they're not ready, as it's used for cluster joining. However, vault, vault-active, and vault-standby services should be able to set
publishNotReadyAddresses
to false to ensure client requests are always routed to a ready pod. Before this PR, all services used the same values.yaml setting, so it wasn't possible to have different values for the different services. Ideally, it would be nice to set the defaultpublishNotReadyAddresses
value to be false at some point, but I just wanted to keep this PR to the minimum possible (minor) breaking change.