Set up an Elasticsearch or OpenSearch target

On the source PostgreSQL database, run:

CREATE SCHEMA shop;
CREATE TABLE shop.orders (
  id integer PRIMARY KEY, customer varchar(100) NOT NULL, status varchar(20),
  amount numeric(12,2), quantity integer, paid boolean,
  created_at timestamp, shipped_at timestamptz, ship_date date,
  note text, meta jsonb, tags text[]);
INSERT INTO shop.orders VALUES
 (1,'Acme Spa','new',1500.50,3,true,'2026-11-01 10:00:00',
  '2026-11-02 08:30:00+01','2026-11-02','first order','{"tier":"gold"}','{a,b}'),
 (2,'Globex','paid',99.99,1,false,'2026-11-01 11:15:00',NULL,NULL,NULL,NULL,NULL);

Create the role. It covers the indexes whose names start with shop-, and the index name orders, because the entity picker reads the mapping of an index named like the source table before you choose the target name:

PUT _security/role/gluesync_writer
{
  "cluster": ["monitor"],
  "indices": [{
    "names": ["shop-*", "orders"],
    "privileges": ["create_index", "write", "delete", "view_index_metadata", "read", "monitor"]
  }]
}

Create a user for Gluesync:

PUT _security/user/gluesync
{ "password": "<a strong password>", "roles": ["gluesync_writer"] }

Or create an API key instead of a user:

POST _security/api_key
{
  "name": "gluesync",
  "role_descriptors": {
    "gluesync_writer": {
      "cluster": ["monitor"],
      "indices": [{
        "names": ["shop-*", "orders"],
        "privileges": ["create_index", "write", "delete", "view_index_metadata", "read", "monitor"]
      }]
    }
  }
}

Open Data pipelines;

Select the + button (Create pipeline), or Create your first pipeline on a new installation. The Agents setup wizard opens on 1. Setup source;

Select the PostgreSQL source and enter its connection, then select Test and continue and Next;

On 2. Setup target(s), search for elastic and select the Elasticsearch agent tile under NoSQL. The form that opens is titled Elasticsearch / OpenSearch;

Enter Node address (a node of the cluster) and Port (9200);

For a user, enter Username and Password. For an API key, leave Username empty, paste the encoded value in Password and set Authentication mode to API_KEY. The other values of Authentication mode are BASIC (the default) and BEARER. Turn on Disable auth only for a cluster that allows anonymous access;

Turn on Enable TLS when the cluster answers on HTTPS. Under Upload 'certificate store' file, select Upload files or Use a path when the cluster uses its own certificate authority;

Optional: set Server flavor to ELASTICSEARCH or OPENSEARCH. The default AUTO recognized both clusters in our test. Under Advanced connection properties, Add property offers Path prefix, Connect timeout (seconds), Request timeout (seconds) and Strict host name check. Additional nodes takes more nodes of the same cluster;

Select Test and continue, then Next;

On 3. Ready, enter a Pipeline name and select Finish.

In the pipeline, select Entities. The Add entities page opens;

In the Object browser, expand the schema shop and tick orders;

Select Confirm and continue. The entity shop.orders is in the Hold group;

Optional: open the three-dot menu of the entity and select Edit mapping. The Field mapping dialog fills Target Index namespace with shop and Target Index with orders, links every field, and shows the field types that Gluesync will create. Select Confirm, then Confirm and continue.

In the row of the entity, select the start icon (Start). The dialog Sync changes (CDC) opens;

In Start from, keep Full copy (marked Recommended for the first start) and keep Keep and merge under Data already in the target;

Select Start entity.

Count and read the documents:

curl -s 'http://<node>:9200/shop-orders/_count'
curl -s 'http://<node>:9200/shop-orders/_search?sort=id&pretty'

Read the mapping that Gluesync created:

curl -s 'http://<node>:9200/shop-orders/_mapping?pretty'

Change the source and read again. These statements update, insert and delete a row:

UPDATE shop.orders SET status = 'shipped', amount = 1600.75 WHERE id = 2;
INSERT INTO shop.orders (id, customer, status, amount, quantity, paid, created_at)
  VALUES (3, 'Initech', 'new', 10.10, 2, false, now());
DELETE FROM shop.orders WHERE id = 1;

Open the gear menu of the pipeline toolbar, select Edit target and open the Advanced settings tab;

Select Add property and pick a key from Key. The keys are Index prefix, Index name template, Insert action, Refresh policy, Truncate strategy, Create index if missing, Explicit mapping, Number of shards, Number of replicas, Document id field, Decimals as strings, Bulk batch size and Bulk request size limit (bytes);

To add a prefix, set Index prefix (for example Acme Shop). The index becomes acme_shop-shop-orders: the name is lower-cased and the space becomes an underscore;

On a one-node cluster, set Number of replicas to 0. The default is 1 shard and 1 replica, and a replica has no node to live on, so the index stays yellow;

To write bigger batches, raise Bulk batch size (default 1000) or Bulk request size limit (bytes) (default 5242880);

Optional: Refresh policy is NONE (default), WAIT_FOR or IMMEDIATE, Insert action is INDEX (default) or CREATE, and Truncate strategy is DELETE_BY_QUERY (default) or SKIP;

Select Test and continue.