Resolve collection failures
Internal product research scaffold for “Resolve collection failures”. Questions and product/code evidence needed before public copy can be written.
Working research scaffold. This page lists questions for Product and Engineering; it does not confirm that the feature is available. The current foundation marks detailed first-party collection, proof-pack generation, buyer inspection, and AI-assisted response preparation as planned. Verify each answer against the released Mockup product and its code before turning it into a public capability claim.
Questions to answer
- Which failure states can occur, and can users tell authentication, permissions, rate limits, network errors, and unsupported data apart?
- Which failures retry automatically, how often, and when do retries stop?
- What action can the user take, and what requires support?
- How does the product avoid exposing credentials or sensitive source data in error details?
- How are partial results represented while a failure is unresolved?
Product and code information needed
- Error types, retry/backoff rules, queue behavior, and dead-letter or terminal-failure handling.
- Customer-facing status and remediation UI, including what each message asks the user to do.
- Redaction rules and logging practices for credentials, API responses, and personal data.
- Monitoring, alert ownership, and tests for retry, recovery, and partial collection.