A huge log table — how do you partition it?

I choose the key to match the queries and retention. For logs, a time range is the natural candidate; region can be a second level if reads and data distribution justify it.

CREATE TABLE logs (
  created_at timestamptz NOT NULL,
  region text,
  message text
) PARTITION BY RANGE (created_at);

CREATE TABLE logs_2026_01 PARTITION OF logs
FOR VALUES FROM ('2026-01-01 00:00:00+00')
TO ('2026-02-01 00:00:00+00');

I check: whether a time filter lets the planner skip partitions, and how we create future partitions and drop old ones.

Pitfall: partitioning within one database is not sharding across many servers. Indexes and the number of partitions still need design. The example has no partitions for other months.