How many simultaneous connections did one instance hold and how did they scale?

The number of simultaneous connections that a single ERP system instance can handle depends on its architecture and the technologies used. In large systems, this can range from a few hundred to several thousand connections.

Scaling is usually done horizontally — by adding new instances and load balancing between them. Optimization methods such as caching, asynchronous request processing, and database optimization are also employed.

For example, clustering servers and using distributed task queues can be used for scaling to ensure high availability and performance as the number of users grows.