» High Availability

Vault is primarily used in production environments to manage secrets. As a result, any downtime of the Vault service can affect downstream clients. Vault is designed to support a highly available deploy to ensure a machine or process failure is minimally disruptive.

» Design Overview

The primary design goal in making Vault highly available (HA) was to minimize downtime and not horizontal scalability. Vault is typically bound by the IO limits of the storage backend rather than the compute requirements. This simplifies the HA approach and allows more complex coordination to be avoided.

Certain storage backends, such as Consul, provide additional coordination functions that enable Vault to run in an HA configuration. When supported by the backend, Vault will automatically run in HA mode without additional configuration.

When running in HA mode, Vault servers have two additional states they can be in: standby and active. For multiple Vault servers sharing a storage backend, only a single instance will be active at any time while all other instances are hot standbys.

The active server operates in a standard fashion and processes all requests. The standby servers do not process requests, and instead redirect to the active Vault. Meanwhile, if the active server is sealed, fails, or loses network connectivity then one of the standbys will take over and become the active instance.

It is important to note that only unsealed servers act as a standby. If a server is still in the sealed state, then it cannot act as a standby as it would be unable to serve any requests should the active server fail.

» Performance Standby Nodes (Enterprise)

Performance Standby Nodes are just like traditional High Availability standby nodes but they can service read-only requests from users or applications. Read-only requests are requests that do not modify Vault's storage. This allows for Vault to quickly scale its ability to service these kinds of operations, providing near-linear request-per-second scaling in many common scenarios for some secrets engines like K/V and Transit. By spreading traffic across performance standby nodes, clients can scale these IOPS horizontally to handle extremely high traffic workloads.

If a request comes into a Performance Standby Node that causes a storage write the request will be forwarded onto the active server. If the request is read-only the request will be serviced locally on the Performance Standby.

Just like traditional HA standbys if the active node is sealed, fails, or loses newtwork connectivity then a performance standby can take over and become the active instance.