<p>We consider customers’ strategic behavior in an <i>N</i>-policy vacation queue with varying arrival and service rates. The arrival and service rates, controlled by the server, vary with queue length and are independent of the server status. Once the system becomes empty, the server may switch off and will only reactivate when the number of customers in the system exceeds a predefined queue length. We first derive the expressions for the mean sojourn time under finite and infinite personal optimal thresholds (i.e., the actual system capacity determined by customers) and analyze their monotonicity. Then, we provide personal optimal algorithms for determining the threshold and explain their rationality under both the fully observable case and the partially observable case. Finally, we present numerical examples to demonstrate these results.</p>