We're going to create a cluster with 2 worker nodes and see the distribution one query on multiple tasks

Open 4 terminals

in the zeroth

docker network create citus-net

in the first

docker run --rm --name citus-1 --network citus-net --hostname citus-1 -p 5501:5432 -e POSTGRES_PASSWORD=postgresqlPassword citusdata/citus:12.1.6-alpine

in the second

docker run --rm --name citus-2 --network citus-net --hostname citus-2 -p 5502:5432 -e POSTGRES_PASSWORD=postgresqlPassword citusdata/citus:12.1.6-alpine

in the third

docker run --rm --name citus-3 --network citus-net --hostname citus-3 -p 5503:5432 -e POSTGRES_PASSWORD=postgresqlPassword citusdata/citus:12.1.6-alpine

to have ability to restart the cluster you need

in the all three nodes (see here)

INSERT INTO pg_dist_authinfo(nodeid, rolename, authinfo) VALUES
    (0, 'postgres', 'password=postgresqlPassword');

in the zeroth

docker exec -it citus-1 psql -U postgres
select * from citus_add_node('citus-2', 5432);
select * from citus_add_node('citus-3', 5432);
select citus_set_coordinator_host('citus-1', 5432);

# https://github.com/citusdata/citus?tab=readme-ov-file#creating-distributed-tables
CREATE TABLE events (
  device_id bigint,
  event_id bigserial,
  event_time timestamptz default now(),
  data jsonb not null,
  PRIMARY KEY (device_id, event_id)
);
SELECT create_distributed_table('events', 'device_id');

SELECT * FROM citus_shards;
SELECT * from pg_dist_shard;
SELECT * from citus_get_active_worker_nodes();
SELECT * FROM citus_check_cluster_node_health();

INSERT INTO events (device_id, data)
SELECT s % 100, ('{"measurement":'||random()||'}')::jsonb FROM generate_series(1,2000000) s;

# https://docs.citusdata.com/en/stable/performance/performance_tuning.html
SET citus.explain_all_tasks = 1;

EXPLAIN (VERBOSE ON) SELECT count(*) FROM events;

you'll get

                                                QUERY PLAN                                                 
-----------------------------------------------------------------------------------------------------------
 Aggregate  (cost=250.00..250.02 rows=1 width=8)
   Output: COALESCE((pg_catalog.sum(remote_scan.count))::bigint, '0'::bigint)
   ->  Custom Scan (Citus Adaptive)  (cost=0.00..0.00 rows=100000 width=8)
         Output: remote_scan.count
         Task Count: 32
         Tasks Shown: All
         ->  Task
               Query: SELECT count(*) AS count FROM public.events_102008 events WHERE true
               Node: host=citus-2 port=5432 dbname=postgres
               ->  Aggregate  (cost=1449.00..1449.01 rows=1 width=8)
                     Output: count(*)
                     ->  Seq Scan on public.events_102008 events  (cost=0.00..1299.00 rows=60000 width=0)
                           Output: device_id, event_id, event_time, data
         ->  Task
               Query: SELECT count(*) AS count FROM public.events_102009 events WHERE true
               Node: host=citus-3 port=5432 dbname=postgres
               ->  Aggregate  (cost=2454.00..2454.01 rows=1 width=8)
                     Output: count(*)
                     ->  Seq Scan on public.events_102009 events  (cost=0.00..2204.00 rows=100000 width=0)
                           Output: device_id, event_id, event_time, data
         ...
(230 rows)

Links