Databases

​

Supported databases

Sequin works with any PostgreSQL database version 14 or higher that supports Logical Replication.

You can find guides for connecting to local and hosted databases in the connect postgres guide.

​

Database connection details

Sequin requires the following details to connect to your database:

• Host: The hostname of your database
• Port: The port number of your database
• Database: The name of your database
• Username: The username of a database user with the necessary permissions
• Password: The password for the database user

You can also enable or disable SSL mode for your database connection.

​

Tunneling to your database

Use the Sequin CLI to create a secure tunnel between Sequin Cloud and a database running on your local machine.

If your database is hosted by a cloud provider and requires a VPN to connect, Sequin can create a secure tunnel through a bastion host. Please contact support to enable this.

​

Database user permissions

We recommend creating a dedicated database user for Sequin. Grant that user the following permissions:

• connect permission on the database.
• select permission on all the tables or schemas you want to connect to Sequin.
• replication permission (to read from replication slots).

You can create a user with those permissions with the following commands:

-- Create user with a secure password
create user sequin_user with password 'REPLACE_WITH_SECURE_PASSWORD';
-- Grant connect permission
grant connect on database your_database to sequin_user;
-- Grant select permission on tables you want to replicate
grant select on table table1, table2, table3 to sequin_user;
-- OR grant select on all tables in a schema
grant select on all tables in schema public to sequin_user;
-- Grant replication permission
alter user sequin_user with replication;

Please create a strong password for Sequin’s database user. You can generate a secure password using the following command:

openssl rand -base64 24

​

Enabling logical replication

Sequin requires logical replication to be enabled on your database. You can check if it’s enabled:

show wal_level;

If the value isn’t logical, you’ll need to enable it. The process for enabling logical replication varies by provider. You can find more information in the following guides:

• Standard PostgreSQL: Modify postgresql.conf to set wal_level = logical
• AWS RDS: Create a parameter group with rds.logical_replication = 1
• Google Cloud SQL: Enable through the console or API
• Azure: Configure through server parameters
• Neon: Enable through project settings
• Supabase: Enabled by default

​

Publications

Publications identify which tables to track for changes.

​

Creating a publication

The simplest publication includes all tables:

create publication sequin_publication for all tables;

To specify a subset of tables, use the following command:

create publication sequin_publication for table table1, table2, table3;

Or all tables in a schema:

create publication sequin_publication for tables in schema public;

​

Working with partitioned tables

When using partitioned tables, you must ensure publish_via_partition_root is set to true to ensure Sequin captures changes correctly. This means changes to partition tables are associated with the root table:

-- Create publication for a partitioned table
create publication sequin_publication for table orders
with (publish_via_partition_root = true);

-- Or modify an existing publication
alter publication sequin_publication set (publish_via_partition_root = true);

SELECT pubviaroot FROM pg_publication where pubname = 'sequin_pub';

​

Modifying Publications

To add or remove tables from a publication, use the ALTER PUBLICATION statement:

-- To remove specific tables
alter publication sequin_publication drop table table1, table2, table3;
-- Add tables
alter publication sequin_publication add table table1, table2, table3;
-- Add all tables in a schema
alter publication sequin_publication add tables in schema my_schema;
-- Drop all tables in a schema
alter publication sequin_publication drop tables in schema my_schema;

​

Deleting a Publication

You can delete a publication using the DROP PUBLICATION statement:

drop publication sequin_publication;

​

Replication slots

Replication slots ensure that PostgreSQL retains WAL (Write-Ahead Log) data needed by Sequin.

​

Creating a Replication Slot

-- Create a logical replication slot that uses the built-in `pgoutput` plugin
SELECT pg_create_logical_replication_slot('sequin_slot', 'pgoutput');

​

Deleting a Replication Slot

SELECT pg_drop_replication_slot('sequin_slot');

​

Best practices

Monitor slot lag Excessive lag in a replication slot can cause WAL buildup and disk space issues. Monitor lag with:

select * from pg_replication_slots;

Clean up unused slots: Delete slots that are no longer in use to prevent WAL retention:

select pg_drop_replication_slot('sequin_slot');

​

Replica identity

Sequin uses REPLICA IDENTITY for two purposes:

1. To determine the old values of an updated row. These values are used to create the changes field for change messages.
2. To load unchanged TOAST columns for updates of both change and row sink consumers.

​

Changes

With REPLICA IDENTITY set to FULL, Sequin receives new and old values for all columns during updates. For columns that have changed values, Sequin includes the old values in the changes field.

​

TOAST columns

PostgreSQL uses TOAST (The Oversized-Attribute Storage Technique) to store large column values. Here’s how Sequin handles TOAST columns:

• TOAST columns are always included for INSERT operations
• For UPDATE operations:
  • Changed TOAST columns are included in rows and changes
  • Unchanged TOAST columns will appear as "unchanged_toast" unless REPLICA IDENTITY is set to FULL
  • If REPLICA IDENTITY is FULL, all TOAST columns are included regardless of changes
• For DELETE operations:
  • TOAST columns are included only if REPLICA IDENTITY is set to FULL

This behavior applies to both changes and rows in Sequin.

To set REPLICA IDENTITY to FULL for a table:

ALTER TABLE your_table REPLICA IDENTITY FULL;

​

Using Sequin with a replica

Sequin supports connecting to a replica database, however in this case Sequin also needs to connect to the primary database.

The only operation that Sequin will perform on the primary database is to call the function pg_logical_emit_message to ensure proper WAL checkpointing.

The only permission required for this is CONNECT on the primary database.

​

PostgreSQL 12 and 13

Sequin relies on the pg_logical_emit_message function for heartbeat messages and backfills. This function is included in Postgres 14 and later.

For Postgres 12 and 13, we support a table-based replacement: public.sequin_logical_messages.

You can create this table in your database with the following SQL:

CREATE TABLE IF NOT EXISTS public.sequin_logical_messages (
  id SERIAL PRIMARY KEY,
  slot_id TEXT NOT NULL,
  subject TEXT NOT NULL,
  content JSONB NOT NULL,
  created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
  updated_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
  UNIQUE(slot_id, subject)
);

-- Grant permissions to Sequin
GRANT SELECT, INSERT, UPDATE, DELETE ON TABLE public.sequin_logical_messages TO sequin_user;
GRANT USAGE ON SEQUENCE sequin_logical_messages_id_seq TO sequin_user;

This table is used by Sequin to replace the pg_logical_emit_message function. It experiences very low write throughput and will not accumulate significant storage or IOPS.

This table must be included in your publication so that Sequin receives messages written to it. If you have created your publication with ALL TABLES, then this is done automatically. Otherwise, you can add the table to your publication with the following SQL:

ALTER PUBLICATION sequin_publication ADD TABLE public.sequin_logical_messages;

​

Backfills

For Postgres 12 and 13, we do not support backfills, which rely heavily on the pg_logical_emit_message function.

If you require backfills, we recommend upgrading to Postgres 14 or later. If this is not an option, please contact us to discuss alternative approaches.

​

Monitoring

Sequin will run health checks against your database connection automatically. You can view the status of the connection at any time in your database’s “Overview” page.