max_parallel_workers
No restart — settable per sessiontuned by pgconfiguratorcount (workers)
The cluster-wide pool of parallel workers all queries draw from.
What pgconfigurator would set
32 GB · 8 vCPU · NVMe · OLTP
Computing…
Tune for your exact server → /tunecomputed in your browser · nothing uploaded
What it does
max_parallel_workers caps the total number of parallel workers active across the whole server at once. Every query's Gather draws from this shared pool (bounded further per query by max_parallel_workers_per_gather), and the pool itself can't exceed max_worker_processes.
How to tune it
Size it to your CPU cores and how much concurrency you expect: too small and queries that planned for workers get fewer (Workers Launched < Planned); too large and parallel queries oversubscribe the CPUs. A common starting point is around the core count.
What the analyzer flags
- Parallel under-use — fewer workers launched than planned