pg_partman with PG12 native partitioning

partition.sql

-- Setup users table and populate
CREATE TABLE users (
  id    INT GENERATED ALWAYS AS IDENTITY PRIMARY KEY,
  email VARCHAR(40) NOT NULL UNIQUE,
  created_at    TIMESTAMP NOT NULL
);
INSERT INTO users (email, created_at)
SELECT
  'user_' || seq || '@' || (
    CASE (RANDOM() * 2)::INT
      WHEN 0 THEN 'gmail'
      WHEN 1 THEN 'hotmail'
      WHEN 2 THEN 'yahoo'
    END
  ) || '.com' AS email,
  timestamp '2016-01-10 20:00:00' +
       random() * (timestamp '2020-01-20 20:00:00' -
                   timestamp '2016-01-10 20:00:00')
FROM GENERATE_SERIES(1, 500000) seq;
CREATE INDEX users_created_at_idx ON users(created_at);
VACUUM ANALYZE users;

-- Use template table so we can have local unique/PK on each partition
CREATE TABLE users_part_template
  (LIKE users INCLUDING ALL);

-- Downtime would effectively start now
ALTER TABLE users RENAME TO users_source; 
CREATE TABLE users
  (LIKE users_part_template INCLUDING DEFAULTS INCLUDING CONSTRAINTS)
PARTITION BY RANGE (created_at);

SELECT partman.create_parent('public.users', 'created_at', 'native', 'monthly', p_premake := 2, p_template_table := 'public.users_part_template');
DROP TABLE users_default;
ALTER TABLE users_source RENAME TO users_default; 
ALTER TABLE users ATTACH PARTITION users_default DEFAULT;
-- Downtime could effectively end now?

-- Move the data from the default partition out to the monthly partitions
CALL partman.partition_data_proc('public.users', '1 month');
VACUUM ANALYZE users;
-- Reindex the original table's indexes to resize them to 0
REINDEX TABLE users_default;

Hey Don, how did things go? Were you able to partition your data with reduced downtime as you planned?

I haven't done this with any actual app databases yet. Thanks for asking!