OpenBao is the open-source fork of HashiCorp Vault. It speaks the same API, so your applications keep working. We move your secrets, policies and login methods, re-point every client and test it — with a rollback ready at every step.
Vault moved to the Business Source License, and HashiCorp is now part of IBM.
Move secrets, policies and logins to OpenBao — in place or side by side — and re-point every client.
Same API for your apps, an open-source licence, rollback ready at every step.
This service is for teams that run Vault today and want their secrets platform on an open-source licence.
OpenBao's own upgrade guide was tested on open-source Vault 1.14.1 with integrated (Raft) storage. Everything else needs the side-by-side route. We tell you which one applies in week one.
| Route | When it fits | How it works |
|---|---|---|
| In-place upgrade | Open-source Vault 1.14.x, integrated (Raft) storage, no external plugins | Snapshot first, then replace Vault with OpenBao node by node; the cluster keeps serving. Fastest route. |
| Side-by-side | Vault Enterprise, Vault 1.15 or newer, Consul storage, external plugins, or a cluster you want to clean up anyway | New OpenBao cluster, configuration kept as code, secrets copied over the API, clients switched in batches. Both clusters run until the last client moves. |
A running OpenBao cluster your team understands, and a Vault you can switch off.
Every secrets engine, login method, policy and client listed. Anything OpenBao handles differently is flagged before work starts.
Highly available, with backups and a tested restore. Configuration kept in Git, so it can be rebuilt.
Applications, CI jobs and Kubernetes workloads read from OpenBao. Each batch is checked before the next one moves.
Unseal method agreed and tested, recovery keys handed over to named people on your side.
How to add a secret engine, rotate a key, restore from backup, upgrade OpenBao. Written for your engineers, not for us.
One live session with your team, then Vault is decommissioned when you say so.
The order is always the same. How long each step takes depends on your volume — the measurement in step 1 sets the price and the dates.
We list mounts, login methods, policies and clients, check versions and plugins, and choose in-place or side-by-side. You get the plan and the final fixed price.
Backup taken and restore tested. OpenBao set up and rehearsed on a copy, so the real cut-over has no surprises.
In-place: node by node. Side-by-side: secrets copied, clients switched in batches, lowest risk first. Rollback ready at every step.
Runbook, live session, recovery keys to your people. Vault is switched off once every client is confirmed on OpenBao.
Fixed price, agreed after the measurement in step 1. Paid by milestone: inventory, build, migration, handover.
Usually not. OpenBao keeps the Vault API, so most clients only need a new address. Things that can break: tools that check the old token format, and plugins OpenBao does not ship. We find these in week one.
Yes, using the side-by-side route: a new OpenBao cluster, configuration and secrets copied over, clients switched one by one. OpenBao's in-place upgrade guide only covers the community edition, so we don't take shortcuts with Enterprise.
Not for your applications if we plan it well. In-place, nodes are upgraded one at a time while the cluster keeps serving. Side-by-side, both clusters run in parallel and clients move in batches.
Yes. Namespaces arrived in OpenBao 2.3 in 2025. If you use Vault Enterprise namespaces, we map them one to one in the migration plan.
OpenBao supports auto-unseal. Which method fits you depends on your KMS or HSM; we confirm it in week one, before you pay for the migration itself.
Version, storage and how many services use it — that's enough for a price range on a 30-minute call.