Troubleshooting

Use this guide when a workspace is connected but not producing useful signals, or when destinations are not receiving useful payloads.

No events are showing up

Check the source integration first:

1. Open Settings > Integrations.
2. Confirm the provider is not disabled.
3. Confirm the provider has an active credential or connection.
4. Copy the endpoint and auth header from the setup guide.
5. Send a small test event.
6. Open Events and check recent rows.

For Direct API and Web SDK setup, use the dedicated installation articles. For PostHog and Segment, remember that provider test buttons may send diagnostic events that do not count as customer evidence.

Events are diagnostic-only

Diagnostic events prove that a source can reach Prevenue, but they usually do not improve signal readiness.

Common causes:

• PostHog forwarded a system event such as $pageview.
• The event name starts with a provider-reserved prefix.
• The event has no stable account identity.
• The event is a setup test rather than a customer action.

Send a custom product or revenue event with account identity to move from diagnostics to useful evidence.

Events are rejected

Open the event row and review the rejection reason. Common causes include:

• Missing event name.
• Missing timestamp.
• Missing account, workspace, customer, company, tenant, or group identity.
• Invalid JSON.
• Wrong bearer credential.
• Unsupported source payload shape.
• Secret-like fields or unsafe data.

Fix the source payload before tuning signals.

Events are hard to map

Mapping is harder when one action has many names or one event name means several actions. Clean up the source when possible:

• Use stable event names.
• Avoid generic names such as clicked, submitted, or updated.
• Put details in properties, not in changing event names.
• Confirm the semantic type for important events.
• Ignore low-value or diagnostic events.

Read Event Mapping and Signal Readiness.

Signal readiness is blocked or partial

Readiness usually stays blocked or partial when:

• Requested motions do not have matching semantic event families.
• Account identity is missing.
• Mappings are not confirmed.
• There are too few customer events.
• A destination is needed before the workflow can become active.
• Signal tuning is required for a custom rule.

Use Setup and the mapping controls in Events to resolve gaps before creating more rules.

Signals are noisy

Start with the signal evidence:

1. Open the signal detail page.
2. Check whether the evidence rows represent meaningful customer behavior.
3. Check whether the triggering events are diagnostic, duplicated, stale, or low-value.
4. Review the Funnel eligibility, follow-up criteria, and cooldown.
5. Suppress the signal if it should not route.
6. Tighten the Funnel criteria after confirming the issue.

Noisy signals usually come from broad event mappings, weak event quality, or Funnel delivery that routes too early.

Destination messages are missing context

Check that source events include the properties your receiving team needs:

• Account name or account key.
• Plan or current plan.
• Target plan or feature.
• Value, quantity, usage ratio, or allowance.
• Reason code.
• Provider or integration name.

Do not solve missing context by sending raw text, transcripts, logs, secrets, or full URLs.

Slack tests fail

Confirm:

• The notification team exists.
• The team has members if email routing is enabled.
• Slack is enabled for the team.
• A channel is connected.
• The workspace has permission to manage Slack.
• The setup test succeeds before relying on automated delivery.

Read Slack Destination Setup.

Webhook tests fail

Confirm:

• The destination URL is HTTPS.
• The receiver responds successfully to test payloads.
• The receiver verifies HMAC signatures with the current signing secret.
• The receiver stores idempotency keys to avoid duplicate downstream work.
• The signing secret was not rotated without updating the receiver.

Read Webhook Destination Setup.

Metrics are locked or empty

Metric cards explain why they are unavailable. Common states include missing Stripe, sync pending, sync failed, product events needed, signals needed, unsupported currency, or insufficient history.

Read Results and Billing.

Prediction readiness has gaps

Prediction readiness gaps are data-sufficiency caveats. They tell you which missing dimensions reduce reliability. Fix them by improving source coverage, event identity, mappings, Stripe match rate, or outcome history.

Read Intelligence Learning Evidence.