In the context of load balancing, what does the term persistence refer to?

Load balancer persistence is implemented via mechanisms such as source IP hashing or session cookies injected by the load balancer, binding a client to a specific server. This is critical for stateful applications that store session data locally on a single server rather than in a shared store. Without persistence, subsequent requests could land on a different server that lacks the client's session state, causing application failures.

Associating multiple HTTP requests to a single user for activity tracking describes application-layer session management, not the load balancing concept of persistence.

Storing TLS session keys to enable faster resumption describes TLS session caching, which is a separate SSL/TLS optimization unrelated to load balancer routing decisions.

Sending dummy traffic to keep idle connections open describes TCP or HTTP keepalive behavior, not load balancer persistence.