HTTP interface for replication

The Replication HTTP API is used internally for synchronizing the nodes in distributed ArangoDB setups, as well as by users to control the replication

The replication architecture and components are described in more details in Replication.

The HTTP replication interface serves four main purposes:

• fetch initial data from a server (e.g. for a backup, or for the initial synchronization of data before starting the continuous replication applier)
• querying the state of a Leader
• fetch continuous changes from a Leader (used for incremental synchronization of changes)
• administer the replication applier (starting, stopping, configuring, querying state) on a Follower

Note that if a per-database setup is used (as opposed to server-level replication), then the replication system must be configured individually per database, and replicating the data of multiple databases will require multiple operations.