Automatically detect and recover from MySQL primary server failures by promoting a healthy replica
Visualize and audit the current MySQL replication topology across all servers and datacenters
Reorganize replication relationships between servers without manual SQL commands
Integrate replication failover into automation pipelines via the HTTP API or CLI
Requires a running MySQL replication setup. Orchestrator itself is a Go binary with a config file. Active development has moved to the Percona fork, use that for new deployments.
Orchestrator is a tool for managing MySQL database replication setups, particularly in situations where high availability matters. MySQL replication is a system where one primary database server continuously copies its data to one or more secondary servers, so the secondaries can take over if the primary fails. Orchestrator sits on top of that setup and helps operators understand, reorganize, and recover from failures in it. The tool runs as a background service and continuously maps the current state of the replication topology, meaning it tracks which servers are replicating from which. It provides a web interface with a visual diagram of this topology, making it easier to see the relationships between servers and spot problems. The same functionality is also available through a command-line tool and an HTTP API for automation. When something goes wrong, such as the primary server going down, Orchestrator can detect the failure automatically and either alert the operator or attempt recovery on its own. Recovery involves promoting one of the secondary servers to become the new primary and updating the other secondaries to replicate from it. The tool understands several different replication modes that MySQL supports, so it can handle the restructuring correctly across different configurations. It also tracks which data center each server is in, which helps when deciding how to redistribute replication responsibilities across physical locations. Orchestrator was originally developed at Outbrain in 2014, then maintained at GitHub from 2016 to 2020, and later moved to an open-source organization. The repository at this location has since been archived. Active maintenance has moved to a fork maintained by Percona, and the README directs contributors there. The project is licensed under Apache 2.0.
← openark on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.