Does anyone run Postgres without PgBouncer?

PostgreSQL, commonly known as Postgres, is a powerful open-source relational database management system. However, it struggles with managing large numbers of connections efficiently. This is where PgBouncer, a connection pooler, comes into play. PgBouncer helps to mitigate this issue by allowing multiple clients to share a limited number of connections to the PostgreSQL server.

Despite the widespread use of PgBouncer, it raises the question: how standard is its usage among managed Postgres providers? A detailed analysis of various managed Postgres providers reveals that a majority of them support PgBouncer or have a similar connection pooling mechanism. This widespread support suggests that managed Postgres providers prioritize ease of use and stability for their users.

However, this standardization comes at a cost. Each provider has had to develop their own connection pooling solution, leading to inconsistencies across providers. Users also need to familiarize themselves with the specifics of each pooler, understanding its limitations and trade-offs. This situation may seem analogous to purchasing a car without a windshield.

While it's the user's responsibility to add necessary components, it's reasonable to question why the provider didn't include a basic feature that most vehicles come equipped with. The absence of built-in connection pooling in managed Postgres offerings may be due to ongoing debates about the trade-offs between processes and threads in PostgreSQL.

Despite the significant effort invested in working around this limitation, providing a default connection pooler like PgBouncer could significantly simplify PostgreSQL deployment and management.