SmartConnector Running and Testing Design Best Practices
Audience: Administrators, Developers, Integrators, Solution Architect
Purpose: Explains the pre-launch and ongoing validation practices that help ensure a live run produces the results you expect, including dry running before any live execution, using diff checking to avoid reprocessing unchanged rows, and validating sample payloads before activating a Webhook SmartConnector.
Overview
Configuration review confirms that a SmartConnector is structurally valid, but validating behavior against real data is what confirms it will produce the results you expect. The practices on this page cover the steps to take before a first live run, during recurring imports, and when building and activating Webhook SmartConnectors.
Always Dry Run Before Going Live
Before running a new SmartConnector against production data, or after making changes to an existing SmartConnector, always perform a dry run first. A dry run executes the full workflow, including SQL processing, variable resolution, and load step evaluation, without writing Records to Kizen.
Review the dry run report before executing a live run. The report surfaces mapping errors, unresolved variables, and load step failures at the row level so you can address them before they affect production data.
For more information on dry runs and the run report, see Running a SmartConnector.
Use Diff Checking on Recurring Imports
For SmartConnectors that ingest the same full dataset on a recurring schedule, enable diff checking to skip rows that have not changed since the previous run. Diff checking compares a hash of each row's resolved execution variable values against the hash stored from the previous run. Rows whose hash matches are skipped before reaching the load step, which is most valuable when the majority of rows in a recurring file are unchanged from run to run.
Note that diff checking compares resolved execution variable values, not raw input data. A change in the input file that does not affect the resolved variable values will not register as a diff. If Records are not updating as expected after a successful run, toggling diff checking off and re-running is the standard first troubleshooting step.
For configuration details, see SmartConnector Diff Checking.
Validate Webhook SmartConnectors Before Activating
When building a Webhook SmartConnector, capture sample payloads while the SmartConnector is inactive. Incoming payloads are accepted and queued but not processed while the SmartConnector is inactive, so you can collect representative sample data without writing anything to Kizen. Before activating, cancel any queued executions you do not want processed, then validate the full workflow using a dry run before going live.
For the full sample payload capture workflow, see Webhook SmartConnectors (Topic Coming Soon).
What's Next
Continue to Build with SmartConnectors (Topic Coming Soon) to learn how to build, test, and deploy SmartConnector SQL from a local environment using the sc CLI, the dev package, and GitHub Actions.
Last updated
Was this helpful?