PostgreSQL + Go Fig
Connect your PostgreSQL databases to Go Fig for real-time financial data integration.
PostgreSQL powers many of the application databases mid-market companies depend on, from the production app to internal data marts. Go Fig connects to read replicas (or a primary in a pinch), pulling operational data into the Financial Intelligence Graph so Celeste can join your application data to the GL, CRM, and payments without anyone exporting CSVs from psql.
Key facts
- Versions
- PostgreSQL 11 and newer
- Auth
- Username/password over TLS, optional client certs
- Source
- Read replica recommended
- Sync mode
- Snapshot, incremental, or logical replication CDC
- Network
- IP allowlist, SSH tunnel, or VPC peering
What you can do with PostgreSQL data in Go Fig
Operational Data Access
Pull live data from your production application database into financial reporting without batch ETL hand-rolling.
Cross-System Joins
Join PostgreSQL application data to QuickBooks, Salesforce, and Stripe inside the Financial Intelligence Graph for unit economics and segment analysis.
Change-Data-Capture Reporting
Use logical replication slots to capture row-level changes for near-real-time dashboards and audit-friendly history.
Data available from PostgreSQL
Go Fig extracts and normalizes the following data from your PostgreSQL account:
How to connect PostgreSQL
Create a read-only Go Fig role
On your PostgreSQL instance, create a dedicated role with CONNECT on the database, USAGE on the relevant schemas, and SELECT on the tables and views you want exposed. Avoid granting superuser. For change-data-capture, also grant REPLICATION and create a logical replication slot scoped to those tables.
Open network access safely
Three options: allowlist Go Fig's published egress IPs on your firewall or RDS security group, configure an SSH tunnel through a bastion, or set up VPC peering / PrivateLink for fully private connectivity. We recommend reading from a replica rather than the primary so analytics load never touches production query latency.
Choose snapshot, incremental, or CDC sync
Snapshot is the simplest path for slowly changing tables. Incremental sync uses an updated_at or sequence column for high-volume tables. Logical replication CDC captures every insert, update, and delete on the tables you select, which is the right mode for tables that change rapidly or where audit-friendly history matters.
Schedule cadence and observe replica lag
Default polling is hourly for snapshot and incremental tables; CDC streams continuously. The connector dashboard surfaces replica lag, slot retained WAL size, and per-table row throughput so your DBA can confirm Go Fig is not pinning resources or accumulating WAL behind a stalled slot.
Authentication: Username and password authentication over TLS is the default. A read-only role scoped to the schemas and tables you want to expose is recommended. Client certificate authentication is supported for environments that require mutual TLS, and SCRAM-SHA-256 is honored when configured.
Common Questions About PostgreSQL Integration
Should Go Fig connect to my primary database or a read replica?
Almost always a read replica. Replicas insulate production query latency from analytics load and let you tune the replica independently for read-heavy workloads. We support connecting to the primary if no replica is available, but we recommend creating one (or pointing at an existing one in your managed service like Aurora, Cloud SQL, or RDS) before going live with frequent sync schedules.
Does Go Fig support change-data-capture via logical replication?
Yes. CDC uses PostgreSQL logical replication slots and the pgoutput plugin (or wal2json on older versions). You grant the Go Fig role REPLICATION privilege, create a slot scoped to the tables you want streamed, and configure the publication. CDC captures every insert, update, and delete with low latency, which is the right mode for high-velocity tables and for any reporting that needs audit-friendly history of row changes.
How does Go Fig keep the connection secure?
TLS is required, with certificate validation enforced by default. Client certificate authentication and SCRAM-SHA-256 are supported. For network-level isolation, customers typically use IP allowlisting against Go Fig's published egress IPs, an SSH tunnel through a bastion, or VPC peering / AWS PrivateLink for fully private routing. The connecting role is scoped to read-only on a curated set of schemas, never granted superuser.
What about managed PostgreSQL services like RDS, Aurora, and Cloud SQL?
All major managed PostgreSQL services are supported, including AWS RDS for PostgreSQL, Aurora PostgreSQL-Compatible, Google Cloud SQL for PostgreSQL, and Azure Database for PostgreSQL. Logical replication is supported on all of them, though each requires a small configuration change (rds.logical_replication parameter on RDS / Aurora, cloudsql.logical_decoding flag on Cloud SQL, wal_level on Azure). Setup docs walk through each.
How does Go Fig handle schema changes in PostgreSQL?
New columns are auto-detected on the next sync and appear in Go Fig automatically. Dropped columns and type changes are flagged rather than silently casting, so your reporting does not quietly start producing wrong numbers when an upstream engineering team alters a table. For CDC streams, schema changes are captured via DDL events when supported, and a notification is sent so flows depending on the changed table can be reviewed.
Ready to connect PostgreSQL?
See how your PostgreSQL data looks in Go Fig with a personalized demo.
Book a Demo