Skip to content
datastudyguides

AI-200 · 2.2.6 · Develop AI solutions by using Azure data management services

Connection pooling

Implement connection optimization to improve throughput and minimize latency

Reuse PostgreSQL connections through an app pool or , and account for transaction-pooling behavior when session state matters.

Reuse connections

PostgreSQL connection pooling flow where many application requests share a small reusable pool before reaching Azure Database for PostgreSQL, with PgBouncer on port 6432.

PostgreSQL uses a server process per connection. Opening one for every request adds latency and memory pressure, so a pool keeps a smaller set open and lends them out.

Choose where the pool lives

Azure Database for PostgreSQL offers built-in on port 6432 for supported tiers. An application can also use a library such as psycopg_pool.

returns the server connection after each transaction. Session-level state such as a plain SET is therefore not guaranteed to be present on the next transaction; use SET LOCAL when the setting belongs to that transaction.

Avoid one PgBouncer sidecar per at large scale, because every sidecar can multiply the connections that reach the database.

Key points

  • Built-in : optional, port 6432 on the same host, not on
  • : default mode; use `SET LOCAL` inside the transaction
  • Pool size: one pool per user and database pair
  • App-side pool: `psycopg_pool.ConnectionPool`; `EntraConnection` fetches a token per connection

Click an underlined term for a one-line explanation, or see them all in the glossary .

Exam trap

Raising the connection limit does not fix connection churn: each connection still costs a server process, so reuse them through a pool.

Sources · 3 Microsoft Learn pages Show ▾

Checked on October 4, 2026. Report a problem

Go deeper on Microsoft Learn

Checked against Microsoft Learn on October 4, 2026. Report a problem with this card

How well do you know this?