VegaFlow quickstart
Connect PostgreSQL, prepare compute and run your first QuickFlow before adding a schedule.
Start with one PostgreSQL source stream and a PostgreSQL destination. Create two saved connections, choose a compute environment and run the sync manually before scheduling it.
Before you begin
- An active organization and workspace.
- Permission to create connections, compute environments and QuickFlows, and to use the required secrets.
- Reachable PostgreSQL source and destination systems, with credentials for the selected read and write operations.
- An available node type and enough cluster capacity.
Store credentials
Create or select a Secret Vault and store the required database passwords. Connection configuration uses secret references containing namespace and key. If you need an SSH tunnel, prepare the SSH credential references as well.
See PostgreSQL authentication for configuration examples.
Create and validate connections
Open VegaFlow → Connections → New connection and choose PostgreSQL. Give the source connection a descriptive name, enter its database settings and select the secret references. Validate and save it, then repeat for the destination.
Both connections use postgres for connector ID and type. Keep their connection IDs distinct and grant the flow operator permission to use each one.
Prepare compute
Open VegaFlow → Clusters and choose an available node type. Set minimum nodes to at least one and maximum nodes to at least the minimum. If auto-stop is enabled, set a positive auto-stop time.
Choose auto-scale and auto-resume settings that fit the workload. Save the environment and wait for it to become ready. See compute environments.
Configure the QuickFlow
Open QuickFlows and select the source connection, destination connection and cluster. Discover the source catalog and select one stream with the columns you need.
Use snapshot mode for a complete read. For incremental cursor mode, choose valid cursor fields, ordering and tie-breakers. Enable schema evolution only if the destination should accept source changes. See selection and source modes.
Run and inspect
Save the QuickFlow and start a manual run. Open its run details and follow the parent state and stream attempts to completion. An accepted run request does not mean the sync succeeded.
Confirm that the selected data reached the destination. Resolve connection, selection or capacity failures before retrying.
Add a schedule
After a successful manual run, create a schedule with an explicit time zone. Check its next run time and observe the first scheduled execution. Pause the schedule when you need to stop future launches while retaining its definition.
See schedules and triggers for schedule changes and run operations for diagnosis.