PostgreSQL Config Calculator

What each setting does

• max_connections = 200 — Hard cap on backends; each one reserves memory even when idle.
• shared_buffers = 4GB — PostgreSQL's own page cache. Requires a restart to change.
• effective_cache_size = 12GB — Planner hint for OS + PostgreSQL cache. Allocates nothing.
• maintenance_work_mem = 1GB — Memory for VACUUM, CREATE INDEX and ALTER TABLE ADD FOREIGN KEY.
• checkpoint_completion_target = 0.9 — Spreads checkpoint writes over 90% of the interval.
• wal_buffers = 16MB — Shared buffer for WAL not yet written to disk.
• default_statistics_target = 100 — Sample size for ANALYZE; higher means better plans, slower ANALYZE.
• random_page_cost = 1.1 — Flash storage: random reads cost about the same as sequential ones.
• effective_io_concurrency = 200 — Concurrent I/O requests the planner assumes the storage can serve.
• work_mem = 10MB — Per sort/hash node, per parallel worker — not per query.
• huge_pages = off — Not worth configuring below ~32 GB of RAM.
• min_wal_size = 1GB — Floor for WAL recycling; avoids constant file creation.
• max_wal_size = 4GB — Triggers a checkpoint once this much WAL accumulates.
• max_worker_processes = 4 — Total background workers, including parallel query and extensions.
• max_parallel_workers_per_gather = 2 — Extra workers a single query may recruit for one Gather node.
• max_parallel_workers = 4 — Parallel workers available across the whole cluster.
• max_parallel_maintenance_workers = 2 — Workers for CREATE INDEX and parallel VACUUM.