Census reads data from one or more tables (possibly across different schemata) in your database and publishes it to the corresponding objects in external systems such as Salesforce. To limit the load on your database as well as to other apps' APIs, Census computes a “diff” to determine changes between each update. In order to compute these diffs, Census creates and writes to a set of tables to a private bookkeeping schema (2 or 3 tables for each sync job configured).
We recommend you create a dedicated
CENSUS user account with a strong, unique password. Census uses this account to connect to your Redshift warehiyse. In order for the Census connection to work correctly, the
CENSUS account must have these permissions:
The ability to create the
CENSUS schema and full admin access to all tables within that schema (including creating tables, deleting tables, and reading and writing to all tables).
Read-only access to any tables and views in any schemata that you would like Census to publish to your service destinations.
Redshift permissions are complex and there are many ways to configure access for Census. The script below has been tested with Redshift and is known to work correctly:
-- Give the census user the ability to sign in with a passwordCREATE USER CENSUS WITH PASSWORD '<strong, unique password>';-- Create a private bookkeeping schema where Census can store sync stateCREATE SCHEMA CENSUS;-- Give the census user full access to the bookkeeping schemaGRANT ALL ON SCHEMA CENSUS TO CENSUS;-- Ensure the census user has access to any objects that may have already existed in the bookkeeping schemaGRANT ALL PRIVILEGES ON ALL TABLES IN SCHEMA CENSUS TO CENSUS;-- Let the census user see this schemaGRANT USAGE ON SCHEMA "<your schema>" TO CENSUS;-- Let the census user read all existing tables in this schemaGRANT SELECT ON ALL TABLES IN SCHEMA "<your schema>" TO CENSUS;-- Let the census user read any new tables added to this schemaALTER DEFAULT PRIVILEGES IN SCHEMA "<your schema>" GRANT SELECT ON TABLES TO CENSUS;-- Let the census user execute any existing functions in this schemaGRANT EXECUTE ON ALL FUNCTIONS IN SCHEMA "<your schema>" TO CENSUS;-- Let the census user execute any new functions added to this schemaALTER DEFAULT PRIVILEGES IN SCHEMA "<your schema>" GRANT EXECUTE ON FUNCTIONS TO CENSUS;
If you have multiple schemata that you would like Census to read from, repeat the steps for "<your schema>" for each of them
In Redshift if there are views in your schema that reference tables in other schemata, you will also need to give Census read access to those other schemata.
If you are using Census models to execute stored procedures (this is rare and not recommended for most users) you may also need to give Census access to those procedures
All connections from the Census Data Warehouse Service to your database, as well as connections from your Redshift database to S3, are protected by TLS encryption - Census will refuse to connect to a warehouse that does not support TLS. All Census data stored in S3 is encrypted with AWS Server-Side Encryption (SSE).
Census will always connect to your data warehouse from of these static IP addresses located within AWS:
Redshift by default prevents any external IP address from accessing your data warehouse so if you will need to add these IP addresses to your security groups. For more information, visit AWS Redshift Help Center.
Advanced methods of Redshift deployment include deploying Redshift within an AWS VPC or private subnet and limiting database access to a separate, controlled database proxy which allows external IP address access. In this configuration, you'll need to separately allow Redshift to communicate directly with S3. For more information on this process, you will need to add an S3 VPC endpoint. This is definitely an obscure feature of AWS but this article does a good job of explaining S3 VPC endpoints, why they're needed and how to set one up.
Census optionally allows connecting to Redshift that are only accessible on private/internal networks via SSH tunneling. To do so, you'll need to provide an SSH host server that is visible on the public internet and can connect to the private warehouse, and you'll also need to be able to perform some basic admin actions on that server.
Create a new user account for Census on the SSH host. (This account is separate from the database user account and can have a different username.)
On the Census connections page, create a new connection to Redshift enter the warehouse connection details, and then check the 'Use SSH Tunnel' option as shown below. Fill in the host and port of the SSH host machine along with the name of the user created in the previous step.
Once the connection is created, Census will generate a keypair for SSH authentication which can be accessed from the connections page. Copy the public key to the clipboard from there and add it to the SSH authorized keys file on the SSH host for the user created in the first step. (If, for example, this user is named 'census', the file will be at /home/census/.ssh/authorized_keys. You may need to create this file if it doesn't exist yet.)
If the SSH host restricts IP ranges that can connect to it, add the Census IPs to the allow list.
With these steps complete, you should be able to complete a connection test, indicating that your tunneled connection is ready to be used in syncs.