Review Paddle Billing subscription consistency with a Postgres cache

## Review Paddle subscription drift against a Postgres cache ### How it works Run a manual, read-only comparison of Paddle Billing subscriptions and a Postgres cache. Live mode reads each source twice, waits at least 10 seconds between observations, and confirms only identical discrepancies. Recent updates and overdue scheduled changes are deferred. The output is a downloadable private HTML report. ### Setup 1. Start with demoMode=true in Configure Review. The demo uses fictional observations 15 seconds apart and bypasses external reads and the Wait node. 2. For live reads, configure your own price-to-plan mapping, five-status access policy, field mapping, environment and settling interval in Configure Review, then set demoMode=false. 3. Bind a Paddle HTTP Header Auth credential with subscription.read permission to Read Paddle A and Read Paddle B. Use Authorization: Bearer YOUR_TOKEN inside the credential, never in this workflow JSON. 4. Bind a SELECT-only Postgres credential with TLS to Read App Cache A and Read App Cache B. Use your cache table/view or adapt both SELECT queries. 5. Run manually and download the binary HTML file from Report Output. ### Scope and limits Maximum 200 records per source; Paddle must report has_more=false. Only one subscription item is supported. No customer identifiers are required. Unknown prices or unsupported data block the review. ### Customization and privacy The example access policy, including past_due=true, is illustrative: choose your own policy. Adjust the settling interval to your cache refresh behavior; it is separate from the 10-second Wait. Keep reports private. This workflow does not modify subscriptions, verify payments or establish contractual entitlement.