HCP Vault Dedicated Gets Cluster Disaster Recovery — and a DR Testing Path

Regional disaster recovery for HCP Vault has protected against cloud provider outages. The new cluster DR feature addresses the failure class that regional DR can’t reach: the cluster-specific incident that takes down Vault while the region stays healthy.

What Happened

HashiCorp introduced cluster disaster recovery (cluster DR) as a public preview capability for HCP Vault Dedicated. The feature complements existing regional DR by enabling cluster-level failover independent of regional infrastructure.

The operational model is support-mediated: teams submit failover requests to HashiCorp support with cluster details and a desired failover window. Requests received within 16 hours are routed to the on-call Vault team, who execute the failover and failback operations and validate the recovery workflow. Cluster DR is available on production-tier HCP Vault Dedicated clusters that already have regional DR enabled.

Why It Matters

Vault is a critical control plane. When Vault is down, deployments halt — services can’t fetch secrets, new workloads can’t authenticate, and infrastructure automation stops. Regional DR handles the scenario where the cloud provider’s region fails. Cluster DR handles the scenarios that regional DR explicitly doesn’t: software bugs that corrupt cluster state, operational errors, and security compromise events where the primary cluster needs to be isolated.

The security response use case is the most operationally significant. If a Vault cluster is suspected of compromise, the current response without cluster DR is difficult: you can’t easily promote the regional secondary without triggering a full regional failover, and taking the primary offline for investigation leaves the control plane dark. Cluster DR enables isolation of the suspected cluster and promotion of the secondary specifically for that cluster — a surgical response rather than a region-level operation.

The other capability cluster DR unlocks is proactive DR testing. Regional outages don’t happen on a schedule. Without cluster DR, a team that wants to validate their Vault recovery procedures has to wait for an actual failure or simulate one in a non-production environment. Cluster DR provides a path to test failover and failback on production-tier infrastructure before an actual incident — which is the only way to have confidence that the runbook works when it matters.

For platform teams that have built AI agent identity infrastructure on Vault’s SPIFFE stack, adding cluster DR to their Vault Dedicated setup closes the resilience gap in the identity control plane that agentic workloads depend on continuously.

What This Means for Platform Engineers

The public preview status means the support-mediated failover model is the initial form — self-service cluster DR is likely in the product roadmap but not yet available. For teams where Vault downtime is a P0 incident, the support path may still be faster than the alternatives during a cluster-specific failure.

Three practical actions for HCP Vault Dedicated production customers: confirm regional DR is enabled (cluster DR requires it), review the support request procedure for cluster DR ahead of any incident, and schedule a DR test using the cluster failover capability to validate that the recovery runbook produces a working cluster on the secondary.

Source: HashiCorp Blog

Content created with AI assistance and reviewed for accuracy.