VaultReactive (health) does not recover after some vault server issues.

[stefan-g]

We are using servicediscovery (Netflix/Eureka) and have several services that are using spring-cloud-vault. After some issues with our vault server/network the health check of the spring-cloud-vault does not recover by him self. The servicediscovery admin page noted that vaultReactive is down.

The following error was on the admin panel:
org.springframework.vault.authentication.VaultLoginException: Cannot retrieve VaultToken from authentication chain; nested exception is org.springframework.web.reactive.function.client.WebClientResponseException$ServiceUnavailable: 503 Service Unavailable from POST https://....

Only a restart of the services solved the problem. I think it would be an improvement when the health check of spring-cloud-vault would recover by him self so that no restart of all services is required.

There is an other exception when the service is shutted down:
Invocation of destroy method failed on bean with name 'reactiveVaultSessionManager': org.springframework.vault.authentication.VaultLoginException: Cannot retrieve VaultToken from authentication chain; nested exception is org.springframework.web.reactive.function.client.WebClientResponseException$ServiceUnavailable: 503 Service Unavailable from POST https://....

It seems to me that the connection is somehow broken. The vault server is available because after restart it works fine.

[mp911de]

Without further details and a proper way to reproduce the issue, there's no chance we can diagnose what's happening here. We'd like to request a reproducer to proceed with this issue.