As a reverse proxy for MySQL databases, ProxySQL is being used in various infrastructure sizes, and it is not surprising to see deployments with thousands of nodes running. Such large deployments introduce some interesting challenges because nodes can be initialized or destroyed at any time.

This session will describe the challenges in configuring such large deployments of nodes, the most common external tools to configure ProxySQL, and then focus on improved ProxySQL native clustering solutions that allow auto-discovery, bootstrap, and a distributed decentralized reconfiguration.