Troubleshooting

Start with the exact app, environment, customer, time, and app/SDK version. Reproduce once and record the observed result. Avoid changing several settings at the same time.

Symptom	First checks
Customer is missing	App key, Test/Live selection, actual SDK startup, and search scope
Sign-in succeeds but the app is locked	Effective access state, exact product mapping, and verified purchase owner
Restore completes without access	Verified transactions, current grants/access, and affected-device observations
Paywall does not appear	Direct versus campaign flow, placement, audience, schedule, status, and renderer compatibility
Reports disagree	App selection, UTC dates, currency, metric meaning, freshness, and coverage
Email is not received	Template, recipient, provider setup, delivery outcome, and destination link
Webhook retries	Receiver response, signing check, saved delivery identity, and prior side effects
A page is empty or slow	Visible request error, selected app, date range, and provider/report availability
Agent cannot read or change a feature	Advertised tool, token scope, allowed apps, environment, and saved plan expiry

Inspect only the evidence needed

Use the customer detail, transaction, Support conversation, report, or delivery that corresponds to the failure. Keep technical payloads in the detail/JSON view. Do not copy private keys, access tokens, full customer exports, or unrelated conversations into a support request.

For Agent work, read context first. A missing permission must not be bypassed using dashboard cookies or database access. Create a new plan when configuration changed after the original preview.

Preserve customer data

Do not recommend deleting an app, account, or local files as a generic first step. Establish how data is stored and whether a usable backup exists. A server-side configuration fix is not confirmation that a device has recovered; verify the customer-facing behavior separately.