Answer
Rather than lowering the polling interval on the Default Queue (which affects every job routed through it), the generally recommended approach is to create a separate, dedicated batch queue specifically for the time-sensitive jobs that need faster pickup, and configure just that new queue's execution plan with the shorter 5-10 second interval. This confines the extra polling overhead to only the jobs that actually need it, rather than applying a more frequent poll cycle system-wide — the right interval still depends on server sizing and database load, so treat it as something to tune per environment rather than a fixed universal number.