Read replicas — Data analysis and reporting

With Neon's read replica feature, you can instantly create a dedicated read-only compute instance for running data-intensive analytics or reporting queries. This allows you to avoid disruption or performance degradation on your production database.

A read replica reads data from the same source as your read-write compute instance. There's no data replication, so creating a read replica is a near-instant process. For more information about Neon's read replica architecture, see Read replicas.

Suppose you have a sales table in your production database. The table and data might look something like this:

CREATE TABLE sales (
    id SERIAL PRIMARY KEY,
    product_id INT NOT NULL,
    sale_amount DECIMAL(10,2) NOT NULL,
    sale_date DATE NOT NULL
);

INSERT INTO sales (product_id, sale_amount, sale_date) VALUES
(1, 20.50, '2022-07-24'),
(2, 35.99, '2022-08-24'),
(1, 20.50, '2022-09-24'),
(3, 15.00, '2023-01-24'),
(1, 20.50, '2023-04-24');
...

You want to find the total sale amount for each product in the past year, but due to the large number of products and sales in your database, you know it's a costly query that could impact performance on your production system.

This guide walks you through creating a read replica, connecting to it, running your query, and optionally deleting the read replica when finished.

Create a read replica

Creating a read replica involves adding a read-only compute endpoint to a branch. You can add a read-only compute endpoint to any branch in your Neon project by following these steps:

1. In the Neon Console, select Branches.
2. Select the branch where your database resides.
3. Click Add compute.
4. On the Create Compute Endpoint dialog, select Read-only as the Compute type.
5. Specify the Compute size options. You can configure a fixed size compute with a specific amount of vCPU and RAM (the default) or enable autoscaling by configuring a minimum and maximum compute size using the slider. You can also configure the Suspend compute after a period of inactivity setting, which is the amount of idle time after which your compute automatically suspends due to inactivity. The default setting is 5 minutes.

   note

   The compute size configuration determines the processing power of your database.

6. When you finish making your selections, click Create.

Your read-only compute is provisioned and appears in the Computes section of the Branches page. This is your read replica. The following section describes how to connect to your read replica.

Alternatively, you can create read replicas using the Neon API or Neon CLI.

curl --request POST \
     --url https://console.neon.tech/api/v2/projects/late-bar-27572981/endpoints \
     --header 'Accept: application/json' \
     --header "Authorization: Bearer $NEON_API_KEY" \
     --header 'Content-Type: application/json' \
     --data '
{
  "endpoint": {
    "type": "read_only",
    "branch_id": "br-young-fire-15282225"
  }
}
' | jq

Connect to the read replica

Connecting to a read replica is the same as connecting to any branch, except you connect via a read-only compute endpoint instead of a read-write compute endpoint. The following steps describe how to connect to your read replica with connection details obtained from the Neon Console.

1. On the Neon Dashboard, under Connection Details, select the branch, the database, and the role you want to connect with.

2. Under Compute, select the RO-replica compute endpoint.

3. Select a Database and the Role you want to connect with.

4. Copy the connection string. This is the information you need to connect to the read replica from you client or application.

   The connection string appears similar to the following:

   postgres://alex:AbC123dEf@ep-cool-darkness-123456.us-east-2.aws.neon.tech/dbname

   If you expect a high number of connections, select Pooled connection to add the -pooler flag to the connection string.

   The information in your connection string corresponds to the following connection details:

   • role: alex
   • password:AbC123dEf
   • hostname: ep-cool-darkness-123456.us-east-2.aws.neon.tech
   • database name: dbname. Your database name may differ.

   When you connect to a read-only compute, no write operations are permitted on the connection.

5. Connect to your application from a client such as psql or add the connection details to your application. For example, to connect using psql, issue the following command:

   psql postgres://alex:AbC123dEf@ep-cool-darkness-123456.us-east-2.aws.neon.tech/dbname

Run the analytics query on the read replica

An analytics query on your sales table might look something like this:

SELECT product_id, SUM(sale_amount) as total_sales
FROM sales
WHERE sale_date >= (CURRENT_DATE - INTERVAL '1 year')
GROUP BY product_id;

If you have a lot of products and sales, this query might impact performance on your production system, but running the query on your read replica, which has its own dedicated compute resources, causes no disruption.

Delete the read replica

When you are finished running analytics queries, you can delete the read replica if it's no longer required. Deleting a read replica is a permanent action, but you can quickly create a new read replica when you need one.

To delete a read replica:

1. In the Neon Console, select Branches.
2. Select a branch.
3. Under Computes, find the read-only compute endpoint you want to delete. Read replicas have a RO replica type.
4. Click the compute endpoint menu on the right side of the table, and select Delete.
5. On the confirmation dialog, click Delete.