My setup:
postgresql.conf: max_connections = 100
pgbouncer config:
[databases]
markets_parser = host=localhost port=5432 dbname=markets_parser user=argon
datestyle=ISO connect_query='SELECT 1'
[pgbouncer]
logfile = /var/log/postgresql/pgbouncer.log
listen_addr = 127.0.0.1
listen_port = 6432
unix_socket_dir = /var/run/postgresql
auth_type = trust
auth_file = /etc/pgbouncer/userlist.txt
admin_users = postgres
stats_users = postgres
pool_mode = session
server_reset_query = DISCARD ALL
max_client_conn = 90
default_pool_size = 70
min_pool_size = 20
reserve_pool_size = 10
I use it from django application (not just site application, but also highload multithreading parsing mechanism) with a bunch of django management commands (which runs as a separate processes), and it works fine (though, almost always, not when I have a sudden splash of activity) when I run only one command in time, and raises a lot of 'not enough connections' errors when I run two commands at once.
What to do?
You can safely increase max_client_conn to a higher number. That is the total number of clients that can connect to pgbouncer, not to the PostgreSQL server. It can be as high as you want, for example 2000 or even higher.
That is indeed one of the main purposes of a connection pooler - to hold small amount of connections open to the database, while allowing high number of end client connections to the pooler itself.
The important thing is to set default_pool_size not higher than the max_connections setting in the postgresql.conf file.