Improve Vault leader election handover
Vault sometimes fails to handover leader election to replicas during pod rotation.
This translates into user errors and alerts for oncall.
Last occurrence: https://gitlab.slack.com/archives/C03QC5KNW5N/p1685996996912499
Previous attempt at improvements at gitlab-com/gl-infra/k8s-workloads/gitlab-helmfiles!2084 (merged)
Desired outcome
Investigate why leader election is slow and fails during pod rotation. Make necessary changes to bring leader election time down to seconds instead of minutes.