Consider coordinating Undici Agent timeouts with WatchConfig Relist Interval

[cmwylie19]

Describe what should be investigated or refactored

We should consider coordinating the WatchConfig RelistTimer interval with the Undici agent timeouts

Links to any relevant code

kubernetes-fluent-client/src/fluent/utils.ts

Line 50 in 5c969a1

export function getHTTPSAgent(opts: RequestInit): Dispatcher | undefined {

Additional context

The risk here, is if there is a silent disconnection from Kubernetes API Server - the relist interval does not re-establish the watch connection.