Rate and pause planner
Daily volume in, schedule out. The planner spreads the load across addresses, sets the gap between requests and gives the pause a spread, because perfectly even intervals are one of the cheapest signals a platform can read.
Why the spread matters
A run that fires a request every 4.0 seconds for six hours is easier to spot than one running twice as fast with human looking gaps. Platforms compare intervals across a session, and a fixed period stands out in that comparison immediately. A spread of a third around the base value costs almost nothing in throughput and removes the pattern.
Spreading load instead of raising speed
Doubling the pace on one address doubles the counter that platform keeps for it. Doubling the number of addresses keeps every counter where it was and finishes the same work in the same time. That is the whole argument for a pool: the limit lives per address, so the way through is more addresses at a calm rate. Addresses that rotate on their own are described on the page for automatic address rotation inside the pool.
Uneven days beat flat ones
The hourly chart above deliberately varies. A collection job that produces exactly the same volume every hour of every day is a signature by itself. Shift the start, let some hours run lighter, and skip an hour entirely once in a while.
Recording what actually happened
Keep the share of refusals per thousand requests for each run. Three or four numbers in a row show whether a change helped, and they turn arguments about pacing into arithmetic. Monthly access to a pool suits this way of working, because the settings survive between runs: see access to the pool for a month.