Parallel processing

Within a single task, domains are processed in parallel — several at a time. So the more domains you load into a task, the faster processing goes per domain: the parallel slots stay busy and the pipeline runs at full capacity.

What follows from this

• One large list beats a set of small tasks: inside a task domains run in parallel, and requests to data sources are grouped into batches. In Domain selection the difference is even more noticeable — there tasks run strictly one at a time, and ten small ones will wait for each other in the queue.
• A task with a single domain uses no parallelism at all — it is the slowest way to check, per domain.
• The number of parallel domains depends on the section and current service load, and is managed automatically — there is nothing to configure.