Connecting to Supabase

​

Connect to Supabase with a direct connection

Sequin uses the Write Ahead Log (WAL) to capture changes from your Supabase database. Supabase’s connection pooling does not support the WAL, so you need to connect to your Supabase database using a direct connection.

Here is how to set up a direct connection in Supabase:

Sequin will automatically detect that you’re using a Supabase database and will check that your connection is configured correctly.

​

Create a replication slot in Supabase

With your database connected, you’ll now create the replication slot that Sequin will use to detect changes in your database.

select pg_create_logical_replication_slot('sequin_slot', 'pgoutput');

create publication sequin_pub for table table1, table2, table3;

create publication sequin_pub for all tables;

​

Moving Supabase webhooks to Sequin

Sequin has some helpful features that Supabase webhooks lack:

• Exactly-once processing guarantees ensuring every event is processed once and only once.
• Automatic retries and backoff for failed events.
• Filtering of events with no PL/pgSQL.
• Backfill support to sync historical data.
• Helpful observability and debugging tools.

You can easily move your existing Supabase webhooks to Sequin by creating Sequin HTTP push consumers.

​

Example Supabase webhook to migrate

As an example, here’s a webhook trigger in Supabase for a hypothetical orders table:

• SQL
• Supabase Console

  create trigger new_order
  after insert on public.orders
  for each row
  execute function supabase_functions.http_request ('https://webhook.site/4f1af4fe-ce43-4530-9d75-f0f2abc086f6', 'POST', '{"Content-type":"application/json"}', '{}', '1000')

In this example, a postgres trigger fires anytime a new record is added to the orders table. The trigger then fires a webhook (via pg_net) to send the data to an HTTP endpoint.

When the trigger fires, Supabase sends the data to the endpoint with the following payload:

{
  "type": "INSERT",
  "table": "orders",
  "record": {
    "id": 4,
    "user_id": 4,
    "order_date": "2024-09-03T18:32:48.108729+00:00",
    "total_amount": 154.97
  },
  "schema": "public",
  "old_record": null
}

​

Sequin HTTP push consumer

To create this exact webhook in Sequin, you’ll create an HTTP push consumer on the orders table with no filters. Here’s how to do it:

alter table orders replica identity full;

Now, when you create a new order in Supabase, Sequin will capture the change and send it to your endpoint:

{
  "record": {
    "id": 5,
    "order_date": "2024-09-03T19:41:17Z",
    "total_amount": "154.97",
    "user_id": 5
  },
  "metadata": {
    "consumer": {
      "id": "107adb3d-76b4-40ba-8e9d-587f9871ab5c",
      "name": "new_order"
    },
    "table_name": "orders",
    "table_schema": "public",
    "commit_timestamp": "2024-09-03T19:41:17.650384Z"
  },
  "action": "insert",
  "changes": null
}

​

Triggering a Supabase edge function with Sequin

Sequin can trigger Supabase edge functions by calling the function’s endpoint URL. This is an alternative to triggering Supabase edge functions with Supabase webhooks. Sequin improves on the Supabase workflow by offering more features, like retries for failed events.

supabase functions deploy {your-function-name}

​

Next steps

Sequin is a great complement to Supabase. This guide showed you how to connect Supabase to Sequin and start moving your webhooks and edge function triggers to Sequin. From here, you may want to:

• Refine your webhooks / triggers by using more specific filters. Perhaps you can move logic out of your edge function and into your consumer filters.
• Use Sequin’s Consume API to gain more control over event consumption.
• Use Sequin’s replay and backfill features to run historical data through your edge function.
• Use Sequin’s observability features to monitor and debug your webhooks and triggers.