B.4. Resource Management #
The parameters below control Postgres Pro AXE server resource usage.
Since each session initializes its own pgpro_axe instance, these parameters are per-session. It is recommended to specify lower values for certain parameters when multiple sessions share the same server resources.
-
duckdb.max_memory/duckdb.memory_limit(integer) # The maximum server memory that pgpro_axe can use to process a single analytical query, in megabytes. For example:
4096(4 GB),8192(8 GB), or1024(1 GB). When set to zero, 80% of server memory is used.Default value:
4096(4 GB).Access: Superuser only.
duckdb.threads/duckdb.worker_threads(integer) #The maximum number of threads that pgpro_axe can use to process a single analytical query.
Default value:
-1— the value is equal to the number of CPU cores.Access: Superuser only.
duckdb.max_workers_per_postgres_scan(integer) #The maximum number of Postgres Pro workers that can be used for a single Postgres Pro scan. This parameter is analogous to
max_parallel_workers_per_gather.You can configure the number of pgpro_axe threads that synchronize with these workers using the
duckdb.threads_for_postgres_scanconfiguration parameter.Default value:
2.Access: All roles.
duckdb.threads_for_postgres_scan(integer) #The maximum number of pgpro_axe threads that can synchronize with Postgres Pro workers for a single Postgres Pro scan.
You can configure the number of Postgres Pro workers using the
duckdb.max_workers_per_postgres_scanconfiguration parameter.Default value:
2.Access: All roles.