This section is fully dedicated to the Workflow Engine scalability and deployment.
- Workflow Engine operates in the single-server mode by default, when it is highly recommended that a single Workflow Engine instance is connected to a single database. Therefore, this mode is not suitable for horizontal scaling. The multi-server mode is the solution.
- In the multi-server mode, any number of Workflow Engine instances can be connected to a single database. All these instances are independent; the only thing that unites them in a cluster is the common database (or schema). The servers can be disconnected and connected to the database at any time; for the cluster to keep on working, one server is enough.
- If a server or all of the servers fail during the operation, all of them or part of them stop, then the data of the active processes launched during the shutdown must be restored. The recovery procedure that can be customized is used for that. But, even without customization, one working server is enough to get everything working properly. If you are using the single server, just restart it to recover.
- Moreover, out-of-the-box Workflow Engine supports multi-tenant applications.