Architecture
BindPlane OP Server is made up of several components. The store, measurements database, and event bus. BindPlane can be configured as a single instance or high availability mode. By default, BindPlane server operates in the single instance mode where all components are included in the installation. Optionally, users can operate BindPlane server in high availability mode. This documentation will cover each component and where they fit within both operation modes.
Single Instance
Single instance is suitable for small, medium, and large deployments of BindPlane. This mode is the easiest way to get started as it requires a single Linux server or virtual machine.
Single instance mode is the recommended deployment model for users who prefer simplicity at the expense of fault tolerance. When fault tolerance is required, BindPlane can be configured in high availability mode.
note
When BindPlane server is down due to maintenance or an outage, managed agents will continue to operate like normal. Agents do not depend on BindPlane to be available.
When operating in single instance mode, BindPlane does not depend on third-party systems such as a remote database.
For more details, see the Single Instance documentation.
High Availability
High availability is a great choice for users who can support a more complex deployment to gain fault tolerance. When operating BindPlane in high availability mode, an individual BindPlane server can go offline for maintenance without causing an outage.
The high availability architecture is significantly more complex, as it requires the following supporting systems:
- 2 or more BindPlane OP servers
- Load balancer
- PostgreSQL database
- Dedicated Prometheus deployment
- Event bus
For more details, see the High Availability documentation.